Files
aimsight-labeling/auto/WILDSTRUBEL.jsonl
T

201 lines
155 KiB
JSON
Raw Normal View History

{"id": "WILDSTRUBEL-1", "start": "2026-09-01T00:00Z", "end": "2026-09-07T00:00Z", "family": "Target issue", "description": "The reference R3 is rejected every day, from about 04:00 to 15:00-16:30, and almost never at night: almost all its measurements from 05:00 to 15:00, almost none from 18:00 to 03:00, from 01/09 to 06/09 as since the start of the data (29/08). Its median error_value is markedly above the threshold by day, and markedly below it at night. On its daytime thumbnails (03/09 at 06:00 and 10:00), its spot cannot be seen, while those of R1 and R2 can; at night, under the infrared, it is sharp, smaller than those of R1 and R2 (03/09 at 02:00). The compensation therefore runs every day on five references, and on three as soon as two others are lost (passages in front of R1 and R2, PA_8 or PA1_2 poorly seen in the sun); the entries and exits of R3 do not make the targets jump (median jump of the corr no more than a little above the noise, like the other steps of the day, apart from the passage of 01/09 at 11:00). Cause not established: a reflector too weak for daytime (small, dirty or poorly oriented). Ignore the curve of R3 by day; act on the target (clean it, reorient it or replace it) to give a sixth reference back to the daytime compensation.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-2", "start": "2026-09-01T03:30Z", "end": "2026-09-01T13:15Z", "family": "Thermal displacement", "description": "From sunrise to early afternoon, the camera moves through thermal effect, as every day: its tilt rises (the camera nods down) markedly above its night level at 07:00, goes markedly below that level in late morning (at 11:30) and regains it at 13:15. At sunrise, the pan and the rotation drop a little (the pan at 05:45, below its night level, the rotation at 05:30), then the rotation goes a little above its level (at 10:30); the scale rises from sunrise (markedly at 05:45) and comes back down only in the afternoon, around 15:45. In the image, targets and references move together, as a block: y markedly positive at 05:45, then a little negative at 12:00 (ensemble.py --rigide; the targets move away from the block only barely). The housing temperature rises markedly from 03h to 10h and stays hot at 13h: the tilt rises with the warming of the morning and comes back while the housing is still hot; the scale follows the temperature and comes back down with it. The compensation follows: on the corrected curves, the tilt wave weighs on each target only barely more than the noise; what remains is the daytime agitation and the deviation of PB1_2 in the light (separate labels). Cause given by the expert for this station (label of 27/09): the camera moves through thermal effect, on its lens or its mount; the road does not move with temperature. Normal, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-3", "start": "2026-09-01T03:45Z", "end": "2026-09-01T16:15Z", "family": "Visibility", "description": "By day, from sunrise to sunset: the error_value of all sensors, references included, rises at sunrise then comes back down in the evening; the peak is at sunrise, when the sun enters the field (seen on the dawn images of the other days). The reference R3 is rejected almost all day, as every day (Target issue, separate label), the target PA_9 at sunrise with the references PA_8 and PA1_2; the corrected curves of the targets are more agitated than at night, without a step. The camera moves in the morning and the compensation follows (separate label, thermal movement of the camera). Cause: daytime heat shimmer above the road, as every day on this station; it cannot be seen directly, the images show the scene in full sun, the spots of the reflectors drowned in the background. Overall well compensated: the structure is not moving, nothing to do; the small daytime oscillations are not a movement.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-4", "start": "2026-09-01T04:15Z", "end": "2026-09-01T16:30Z", "family": "Visibility", "description": "PB1_2 alone, by day: its error_value rises much more than that of its neighbours PB2_2 and PB4 and approaches the threshold; its corr_x moves away on one side all day and comes back in the evening. On the daytime thumbnails, the target is placed at the edge of the pavement kerb, bright in full sun, and its spot is hard to make out there; at night, under the infrared, the spot is sharp. Cause deduced from the thumbnails: the daylight on the neighbouring kerb; the same time slot comes back every day. The structure is not moving: do not trust the curve of PB1_2 by day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-5", "start": "2026-09-01T10:45Z", "end": "2026-09-01T11:30Z", "family": "Obstruction", "description": "A car passes in front of the targets PB2_2 and PB_5 (seen on their 11:00 thumbnails) and also hides PA_9 and the reference PB3_2: their error_value jumps very high, their measurements are rejected for the time of one measurement. At the next measurement, lights and a vehicle trail pass near the references R1 and R2 (11:15 thumbnails): R1, R2 and PA_8 are rejected in turn. The other targets do not move; everything comes back at 11:30. The structure is not moving, nothing to do: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-6", "start": "2026-09-01T14:30Z", "end": "2026-09-01T15:00Z", "family": "Obstruction", "description": "PA3_2 alone: for the time of one measurement, a row of bright points appears just above the target (14:45 thumbnail, absent at 14:30 and at 15:00), probably the lights or reflectors of a passing vehicle; the measurement locks onto one of them while staying under the threshold, and corr_x, corr_y of PA3_2 make a very large jump, then come back at the next measurement. The other targets and the compensation do not move. The structure is not moving: this point of PA3_2 is false, to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-7", "start": "2026-09-01T16:15Z", "end": "2026-09-01T16:45Z", "family": "Obstruction", "description": "At nightfall, for the time of one measurement, the spot of the reference R1 disappears (16:30 thumbnail): its error_value jumps very high, it is rejected; R2, R3 and the target PB4 also rise. Without R1, the basis changes: PB2_2, PB4 and PB_5 jump together on corr_x or corr_y, then come back at the next measurement; their raw positions do not move. Probable cause: a vehicle passing in front of R1, which the night image does not show. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-8", "start": "2026-09-02T03:45Z", "end": "2026-09-02T13:15Z", "family": "Thermal displacement", "description": "From sunrise to early afternoon, the camera moves through thermal effect, as every day: its tilt rises (the camera nods down) markedly above its night level at 06:30, goes markedly below that level in late morning (at 11:30) and regains it at 13:15. At sunrise, the pan and the rotation drop a little (the pan at 05:00, below its night level, the rotation at 04:45), then the rotation goes a little above its level (at 10:45); the scale rises from sunrise (markedly at 05:00) and comes back down only in the afternoon. In the image, targets and references move together, as a block: y markedly positive at 06:30, then a little negative at 11:30 (ensemble.py --rigide; the targets move away from the block only barely). The housing temperature rises markedly from 03h to 07h and stays hot at 13h: the tilt rises with the warming of the morning and comes back while the housing is still hot; the scale follows the temperature and comes back down with it. The compensation follows: on the corrected curves, the tilt wave weighs only a little at most (PB1_2 on corr_y); what remains is the daytime agitation and the deviation of PB1_2 in the light (separate labels). Cause given by the expert for this station (label of 27/09): the camera moves through thermal effect, on its lens or its mount; the road does not move with temperature. Normal, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-9", "start": "2026-09-02T03:45Z", "end": "2026-09-02T16:15Z", "family": "Visibility", "description": "By day, from sunrise to sunset: the error_value of all sensors, references included, rises at sunrise then comes back down in the evening; at sunrise, the sun is in the field, with reflections on the lens (04:45 image). The reference R3 is rejected almost all day, as every day (Target issue, separate label), PA_9 at sunrise with PA_8 and PA1_2; at 05:15 the compensation rejects all its references for the time of one measurement and PA3_2 is rejected; at 07:30 and at 11:15, vehicles pass near the references R1 and R2 (separate labels). The corrected curves of the targets are more agitated than at night, without a step; the camera moves in the morning and the compensation follows (separate label, thermal movement of the camera). Cause: daytime heat shimmer above the road, as every day on this station, not seen directly. Overall well compensated: the structure is not moving, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-10", "start": "2026-09-02T04:15Z", "end": "2026-09-02T16:45Z", "family": "Visibility", "description": "PB1_2 alone, by day: its error_value rises much more than that of its neighbours and approaches the threshold; its corr_x moves away on one side all day and comes back in the evening. On the daytime thumbnails (05:00, 09:00, 13:00), the target is at the edge of the pavement kerb, bright in full sun, and its spot is hard to make out there; at night, the spot is sharp. Cause deduced from the thumbnails: the daylight on the neighbouring kerb, every day. The structure is not moving: do not trust the curve of PB1_2 by day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-11", "start": "2026-09-02T07:15Z", "end": "2026-09-02T07:45Z", "family": "Obstruction", "description": "At 07:30, vehicles pass near the references R1 and R2: headlights and bright spots on their 07:30 thumbnails. The compensation rejects them for the time of one measurement, the basis changes, and PB4 and PB_5 jump on corr_x (PB4 also on corr_y); everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-12", "start": "2026-09-02T11:00Z", "end": "2026-09-02T11:30Z", "family": "Obstruction", "description": "At 11:15, the references R1 and R2 are rejected for the time of one measurement: on their 11:15 thumbnails, blurry translucent shapes (an inscription near R1, a bright spot near R2), a passing vehicle. The basis changes, PB4 and PB_5 jump on corr_x; everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-13", "start": "2026-09-02T11:45Z", "end": "2026-09-02T12:15Z", "family": "Obstruction", "description": "A vehicle passes in front of PA3_2 and PA_9 (trails and reflections on their 12:00 thumbnails): their error_value jumps, their measurements are rejected for the time of one measurement, along with the reference R2. The other targets do not move; everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-14", "start": "2026-09-02T12:15Z", "end": "2026-09-02T12:45Z", "family": "Obstruction", "description": "A vehicle passes on the lane along the pavement, seen as translucent (motion blur) on the 12:30 image: it hides PB1_2, PB2_2 and the reference PA1_2, whose error_value jumps and which are rejected for the time of one measurement. The other targets do not move; everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-15", "start": "2026-09-03T02:45Z", "end": "2026-09-06T03:00Z", "family": "Obstruction", "description": "PB1_2 alone: at 03:00 on 03/09, its spot almost disappears (03:00 and 03:15 thumbnails, while that of its neighbour PB2_2 is sharp) and its error_value jumps very high. Then, for three days, its spot stays smaller at night (23:00 thumbnails on 03/09 and 04/09, compared with 02/09) and its night error_value stays above its level; by day, it exceeds the threshold for most of the day (measurements rejected) and its corr_x dips much more than on the other days; the night level of its corr_x drops, not that of its neighbours. On 06/09 at 02:45, a one-measurement spike on corr_x, then the night error_value comes back to its level and the spot becomes whole again (06/09 23:00 thumbnail). Cause: something partly covers the spot of PB1_2, of a nature not seen on the images. The structure is not moving: do not trust the curve of PB1_2 over the range; look at what is in front of the target if it comes back.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-16", "start": "2026-09-03T03:45Z", "end": "2026-09-03T13:45Z", "family": "Thermal displacement", "description": "From sunrise to early afternoon, the camera moves through thermal effect, as every day: its tilt rises (the camera nods down) markedly above its night level at 06:30, goes markedly below that level in late morning (at 11:30) and regains at 13:45 the afternoon level, a little below that of the night. At sunrise, the pan and the rotation drop a little (the pan at 05:00, below its night level, the rotation at 05:00), then the rotation goes a little above its level (at 09:15); the scale rises from sunrise (markedly at 06:00) and comes back down only in the afternoon. In the image, targets and references move together, as a block: y markedly positive at 06:30, then a little negative at 11:30 (ensemble.py --rigide; the targets move away from the block only barely). The housing temperature rises markedly from 03h to 08h and stays hot at 13h: the tilt rises with the warming of the morning and comes back while the housing is still hot; the scale follows the temperature and comes back down with it. The compensation follows: on the corrected curves, the tilt wave weighs only a little at most (PB1_2 and PA_9 on corr_x); what remains is the daytime agitation and the deviation of PB1_2, partly covered that day (separate labels). Cause given by the expert for this station (label of 27/09): the camera moves through thermal effect, on its lens or its mount; the road does not move with temperature. Normal, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-17", "start": "2026-09-03T03:45Z", "end": "2026-09-03T15:15Z", "family": "Visibility", "description": "By day, from sunrise to sunset: the error_value of all sensors, references included, rises at sunrise then comes back down in the afternoon. The reference R3 is rejected almost all day, as every day (Target issue, separate label), PA_9 at sunrise with PA_8 and PA1_2; at 07:30, a vehicle passes near the references R1 and R2 (separate label); at 09:15, R2 is rejected alone for the time of one measurement, with no jump of the targets. The corrected curves of the targets are more agitated than at night, without a step; the camera moves in the morning and the compensation follows (separate label, thermal movement of the camera). Cause: daytime heat shimmer above the road, as every day on this station, not seen directly. Overall well compensated: the structure is not moving, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-18", "start": "2026-09-03T03:45Z", "end": "2026-09-03T15:15Z", "family": "Visibility", "description": "PB1_2 alone, by day: its error_value exceeds the threshold for most of the day (measurements rejected) and its corr_x moves away on one side, more strongly than on the previous days, before slowly coming back in the evening. As every day, the target is at the edge of the pavement kerb, bright in full sun on the daytime thumbnails; that day its spot is also partly covered (separate label, from 03/09 to 06/09). Cause deduced from the thumbnails: the daylight on the neighbouring kerb. The structure is not moving: do not trust the curve of PB1_2 by day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-19", "start": "2026-09-03T07:15Z", "end": "2026-09-03T07:45Z", "family": "Obstruction", "description": "At 07:30, the headlights of a vehicle shine near the references R2 and R1 (07:30 thumbnails): they are rejected for the time of one measurement, the basis changes, and PB2_2, PB4 and PB_5 jump on corr_x, PB4 also on corr_y; everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-20", "start": "2026-09-03T08:00Z", "end": "2026-09-03T08:30Z", "family": "Obstruction", "description": "A large vehicle passes very close to the camera, in front of the whole scene, seen as translucent on the 08:15 image: the references R1, R2, R3, PA1_2 and PA_8 are rejected, the camera is not recalibrated, the targets PA3_2 and PA_9 are rejected (PB1_2 already is all morning, hindered by the daylight) and PB4 makes a small jump on corr_x. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-21", "start": "2026-09-03T21:30Z", "end": "2026-09-03T22:00Z", "family": "Obstruction", "description": "At night, the target PA2_2 alone is hidden for the time of one measurement: its error_value jumps very high, it is rejected. On the 21:45 thumbnails, a large bright spot (a headlight) and a licence plate near the target: a passing vehicle. The other sensors do not move; everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-22", "start": "2026-09-04T03:30Z", "end": "2026-09-04T13:00Z", "family": "Thermal displacement", "description": "From sunrise to early afternoon, the camera moves through thermal effect, as every day: its tilt rises (the camera nods down) markedly above its night level at 06:15, goes markedly below that level in late morning (at 11:00) and regains at 13:00 the afternoon level, a little below that of the night. At sunrise, the pan and the rotation drop a little (the pan at 05:45, below its night level, the rotation at 05:45), then the rotation goes a little above its level (at 10:15); the scale rises from sunrise (markedly at 07:45) and comes back down only in the afternoon, around 15:00. In the image, targets and references move together, as a block: y markedly positive at 07:15, then a little negative at 11:00 (ensemble.py --rigide; the targets move away from the block only barely). The housing temperature rises markedly from 03h to 10h and stays hot at 13h: the tilt rises with the warming of the morning and comes back while the housing is still hot; the scale follows the temperature and comes back down with it. The compensation follows: on the corrected curves, the tilt wave weighs only a little, except on PB1_2 (markedly more on corr_y), hindered by the light and partly covered that day; what remains is the daytime agitation and these episodes of PB1_2 (separate labels). Cause given by the expert for this station (label of 27/09): the camera moves through thermal effect, on its lens or its mount; the road does not move with temperature. Normal, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-23", "start": "2026-09-04T03:45Z", "end": "2026-09-04T15:15Z", "family": "Visibility", "description": "By day, from sunrise to sunset: the error_value of all sensors, references included, rises at sunrise then comes back down in the afternoon. The reference R3 is rejected almost all day, as every day (Target issue, separate label), PA_9 at sunrise with PA_8 and PA1_2; at 05:45 and 06:00, at the height of sunrise, the compensation rejects all its references and PA2_2, PA3_2 and PB_5 make a small jump; at 10:15, a vehicle passes near the references R1 and R2 (separate label). The corrected curves of the targets are more agitated than at night, without a step; the camera moves in the morning and the compensation follows (separate label, thermal movement of the camera). Cause: daytime heat shimmer above the road, as every day on this station, not seen directly. Overall well compensated: the structure is not moving, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-24", "start": "2026-09-04T03:45Z", "end": "2026-09-04T15:30Z", "family": "Visibility", "description": "PB1_2 alone, by day: its error_value exceeds the threshold for a large part of the day (measurements rejected in late morning) and its corr_x moves away on one side, then comes back in the evening. As every day, the target is at the edge of the pavement kerb, bright in full sun on the 09:00 thumbnail; its spot is also partly covered over the range from 03/09 to 06/09 (separate label). Cause deduced from the thumbnails: the daylight on the neighbouring kerb. The structure is not moving: do not trust the curve of PB1_2 by day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-25", "start": "2026-09-04T06:30Z", "end": "2026-09-04T07:00Z", "family": "Obstruction", "description": "A vehicle passes in front of the targets (headlights near PA3_2, light-coloured bodywork behind PA2_2 and PB2_2 on the 06:45 thumbnails): PA_9 and PB_5 are rejected, with the references PA_8 and R2; PA3_2, which stayed under the threshold, locks onto a headlight and its corr_x, corr_y make a very large jump. Everything comes back at the next measurement. The structure is not moving: this point is false, to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-26", "start": "2026-09-04T10:00Z", "end": "2026-09-04T10:30Z", "family": "Obstruction", "description": "At 10:15, a vehicle passes near the references R1 and R2: lights near R1 and translucent shapes near R2 on their 10:15 thumbnails. They are rejected for the time of one measurement, the basis changes, and PB4, PB_5 and PA3_2 jump; everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-27", "start": "2026-09-04T11:30Z", "end": "2026-09-04T12:30Z", "family": "Obstruction", "description": "Like on 02/09 at 12:30, a vehicle passes on the lane along the pavement, seen as translucent on the 12:00 image: PB2_2 and the references PA1_2 and PA_8 are rejected for the time of one measurement; PB1_2, already narrowly rejected by the daylight, sees its error_value rise very high. At the measurement before, the reference R2 is already rejected, a blurry translucent shape on its 11:45 thumbnail; at the measurement after, PA_8 still is. The other targets do not move; everything comes back at 12:30. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-28", "start": "2026-09-04T13:45Z", "end": "2026-09-04T14:15Z", "family": "Obstruction", "description": "A vehicle passes in front of PA3_2, PA_9, PB4 and PB_5 (blurry shapes on their 14:00 thumbnails): their error_value jumps, their measurements are rejected for the time of one measurement. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-29", "start": "2026-09-04T15:00Z", "end": "2026-09-04T15:30Z", "family": "Obstruction", "description": "PB4 alone is rejected for the time of one measurement: its measurement slides toward small bright points that appeared near it on the 15:15 thumbnail, probably the reflections of a passing vehicle (cause not established). The other sensors do not move; everything comes back at the next measurement. Point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-30", "start": "2026-09-05T02:00Z", "end": "2026-09-05T02:30Z", "family": "No data", "description": "The 02:15 measurement is missing for all sensors, the image of that hour too; the metrics continue, the battery and the voltage at their night level. JUNGFRAU and NORDEND also lose a measurement around the same hour on those nights (outside their batches): a common cause outside the station, at night, is probable; it is not established. The curves are missing for the time of one measurement; nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-31", "start": "2026-09-05T03:30Z", "end": "2026-09-05T08:00Z", "family": "Obstruction", "description": "Cars parked along the pavement hide the reference PB3_2 and the targets PA2_2, PA3_2, PB2_2 and PB_5, seen on the images from 03:45 to 07:45; the scene is clear at 08:00. These sensors have a very high error_value and are rejected almost throughout the range; PA2_2, kept at 04:30, makes a very large jump for the time of one measurement (the measurement probably locks onto a reflection on a car). With PB3_2 hidden and, at sunrise, PA_8 and R3 rejected, the compensation runs on few references, then on none at 05:45 and 06:00 (no recalibration at 06:00): PB4 then makes a one-measurement spike on corr_y. Everything comes back at 08:00. The structure is not moving: ignore the curves of these targets and the spike of PB4 over the range.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-32", "start": "2026-09-05T03:30Z", "end": "2026-09-05T12:45Z", "family": "Thermal displacement", "description": "From sunrise to early afternoon, the camera moves through thermal effect, as every day: its tilt rises (the camera nods down) markedly above its night level at 07:15, goes markedly below that level in late morning (at 11:00) and regains it at 12:45. At sunrise, the pan and the rotation drop a little (the pan at 05:45, below its night level, the rotation at 05:45), then the rotation goes a little above its level (at 10:15); the scale rises from sunrise (markedly at 05:45) and comes back down only in the afternoon. In the image, targets and references move together, as a block: y a little positive at 08:00, after the parked cars leave (separate label), then as much negative at 11:00 (ensemble.py --rigide; the targets move away from the block only barely). The housing temperature rises markedly from 03h to 08h and stays hot at 12h: the tilt rises with the warming of the morning and comes back while the housing is still hot; the scale follows the temperature and comes back down with it. The compensation follows: on the corrected curves, the tilt wave weighs on each target only barely more than the noise; what remains is the daytime agitation, the parked cars of the morning and the deviation of PB1_2 (separate labels). Cause given by the expert for this station (label of 27/09): the camera moves through thermal effect, on its lens or its mount; the road does not move with temperature. Normal, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-33", "start": "2026-09-05T03:45Z", "end": "2026-09-05T15:15Z", "family": "Visibility", "description": "By day, from sunrise to sunset: the error_value of all sensors, references included, rises at sunrise, the sun in the field (03:45 image), then comes back down in the afternoon. The reference R3 is rejected almost all day, as every day (Target issue, separate label), PA_9 at sunrise; at 09:15, a vehicle passes in front of PA_9 (separate label); the morning is mixed with the parked cars in front of some of the targets (separate label). The corrected curves of the targets are more agitated than at night, without a step; the camera moves in the morning and the compensation follows (separate label, thermal movement of the camera). Cause: daytime heat shimmer above the road, as every day on this station, not seen directly. Overall well compensated: the structure is not moving, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-34", "start": "2026-09-05T03:45Z", "end": "2026-09-05T16:45Z", "family": "Visibility", "description": "PB1_2 alone, by day: its error_value exceeds the threshold for a large part of the day (measurements rejected from late morning to early afternoon) and its corr_x moves away on one side, then comes back in the evening. As every day, the target is at the edge of the pavement kerb, bright in full sun on the 09:00 thumbnail; its spot is also partly covered over the range from 03/09 to 06/09 (separate label). Cause deduced from the thumbnails: the daylight on the neighbouring kerb. The structure is not moving: do not trust the curve of PB1_2 by day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-35", "start": "2026-09-05T09:00Z", "end": "2026-09-05T09:30Z", "family": "Obstruction", "description": "At 09:15, PA_9 alone is rejected for the time of one measurement: on its 09:15 thumbnail, bright points around the target and a dark shape, the lights of a vehicle seen as translucent. The other targets do not move; everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-36", "start": "2026-09-05T11:45Z", "end": "2026-09-05T12:15Z", "family": "Obstruction", "description": "A vehicle passes in front of PA3_2 and PB4 (blurry shapes on their 12:00 thumbnails): their measurements are rejected for the time of one measurement, with the reference R2, and PA2_2 makes a small jump. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-37", "start": "2026-09-05T23:00Z", "end": "2026-09-06T00:30Z", "family": "Obstruction", "description": "At night, a parked car hides the reference PB3_2 and the targets PA2_2, PA3_2, PB2_2 and PB_5: their spots disappear on the thumbnails, and the car's licence plate shines under the infrared (23:30 image). Their error_value rises very high, their measurements are rejected for an hour; the other targets and the compensation, on the other references, do not move. Everything comes back at 00:30, the car gone. The structure is not moving: ignore the curves of these targets over the range.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-38", "start": "2026-09-06T01:45Z", "end": "2026-09-06T02:15Z", "family": "No data", "description": "The 02:00 measurement is missing for all sensors, the image of that hour too; the metrics continue, the battery and the voltage at their night level. JUNGFRAU and NORDEND also lose a measurement around the same hour on those nights (outside their batches): a common cause outside the station, at night, is probable; it is not established. The curves are missing for the time of one measurement; nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-39", "start": "2026-09-06T03:45Z", "end": "2026-09-06T13:15Z", "family": "Thermal displacement", "description": "From sunrise to early afternoon, the camera moves through thermal effect, as every day: its tilt rises (the camera nods down) markedly above its night level at 06:00, goes markedly below that level in late morning (at 11:45) and regains at 13:15 the afternoon level, a little below that of the night. At sunrise, the pan and the rotation drop a little (the pan at 06:00, below its night level, the rotation at 06:00), then the rotation goes a little above its level (at 11:00); the scale rises from sunrise (markedly at 07:00) and comes back down only in the afternoon, around 15:00. In the image, targets and references move together, as a block: y markedly positive at 05:15, then a little negative at 11:45 (ensemble.py --rigide; the targets move away from the block only barely). The housing temperature rises markedly from 03h to 08h and stays hot at 13h: the tilt rises with the warming of the morning and comes back while the housing is still hot; the scale follows the temperature and comes back down with it. The compensation follows: on the corrected curves, the tilt wave weighs only a little, except on PB_5 (a little more on corr_x, over few measurements only); what remains is the daytime agitation and the deviation of PB1_2 in the light (separate labels). Cause given by the expert for this station (label of 27/09): the camera moves through thermal effect, on its lens or its mount; the road does not move with temperature. Normal, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-40", "start": "2026-09-06T03:45Z", "end": "2026-09-06T14:30Z", "family": "Visibility", "description": "PB1_2 alone, by day: its error_value rises much more than that of its neighbours, exceeds the threshold at times (a few measurements rejected), and its corr_y then its corr_x move away from their level before coming back in the evening. As every day, the target is at the edge of the pavement kerb, bright in full sun on the 09:00 thumbnail. Cause deduced from the thumbnails: the daylight on the neighbouring kerb. The structure is not moving: do not trust the curve of PB1_2 by day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-41", "start": "2026-09-06T03:45Z", "end": "2026-09-06T15:00Z", "family": "Visibility", "description": "By day, from sunrise to sunset: the error_value of all sensors, references included, rises at sunrise then comes back down in the afternoon. The reference R3 is rejected almost all day, as every day (Target issue, separate label), PA_9 at sunrise with PA_8; at 07:15, when the compensation rejects its references for the time of one measurement, PB4, PB2_2 and PB_5 make a small jump; at 09:30, a vehicle passes (separate label). The corrected curves of the targets are more agitated than at night, without a step; the camera moves in the morning and the compensation follows (separate label, thermal movement of the camera). Cause: daytime heat shimmer above the road, as every day on this station, not seen directly. Overall well compensated: the structure is not moving, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-42", "start": "2026-09-06T05:45Z", "end": "2026-09-06T06:15Z", "family": "Obstruction", "description": "At sunrise, a blurry silhouette passes on the pavement near PB1_2 and PB2_2 (06:00 image), which are rejected for the time of one measurement; with the references rejected by the morning light, the compensation retains none at that moment, and PA3_2, PB4 and PB_5 make a small jump on corr_x. Everything comes back at the next measurement. Probable cause: a passer-by; the structure is not moving, point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-43", "start": "2026-09-06T09:15Z", "end": "2026-09-06T09:45Z", "family": "Obstruction", "description": "For the time of one measurement, a vehicle passes in the background, near the references R1 and R2: its lights and bodywork can be seen as translucent on their 09:30 thumbnails. R1, R2 and PA_8 are rejected, the basis changes and PB4 makes a small jump; everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-44", "start": "2026-09-06T12:00Z", "end": "2026-09-06T12:30Z", "family": "Obstruction", "description": "A light-coloured vehicle passes on the lane along the pavement and covers PB1_2 and PB2_2 (12:15 thumbnails): they are rejected for the time of one measurement, with the reference PA_8. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-45", "start": "2026-09-06T14:15Z", "end": "2026-09-06T14:45Z", "family": "Obstruction", "description": "For the time of one measurement, bright points appear near PA_9 and the reference PA_8 (14:30 thumbnail, absent at 14:15 and at 14:45), probably the reflections of a passing vehicle: PA_8 is rejected; PA_9, which stayed under the threshold, locks onto one of them and its corr_x, corr_y make a very large jump. Everything comes back at the next measurement. The structure is not moving: this point of PA_9 is false, to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-46", "start": "2026-09-07T00:00Z", "end": "2026-10-01T00:00Z", "family": "Target issue", "description": "The reference R3 is rejected on every measured day of the batch (from 07 to 09/09, then from 22 to 30/09), from sunrise until mid-afternoon, and almost never at night: by day, its error_value stays above its threshold at almost every measurement, a little less often in the last days of September; at night, it is well below. On its daytime thumbnail (30/09 at 10:00), its spot cannot be seen, while those of R1 and R2 can; at night, under the infrared, it is sharp, smaller than theirs (03/09 at 02:00 thumbnail). It is rejected as much before as after the outage from 09/09 to 22/09, when the other reflectors are better seen: it is the target, not the daytime sun. The compensation therefore runs every day on five references, and on three or fewer as soon as two others are lost (PA_8 poorly seen before the outage, passages in front of R1 and R2, sunrise, separate labels); its entries and exits do not make the targets jump. Cause not established: a reflector too weak for daytime (small, dirty or poorly oriented). Ignore the curve of R3 by day; act on the target (clean it, reorient it or replace it) to give a sixth reference back to the daytime compensation.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-47", "start": "2026-09-07T03:30Z", "end": "2026-09-07T13:15Z", "family": "Thermal displacement", "description": "At sunrise, the camera nods down: its tilt rises, with its scale and its rotation, and all the raw positions, references included, move together in the image; the tilt peaks in early morning, goes below its night level in late morning, then comes back to it in early afternoon, as every day, while the housing temperature rises; the scale and the pan come back more slowly, until mid-afternoon. The compensation follows: the corrected curves of the targets do not show this wave, only the daytime agitation (separate label). Deduced cause: the mount or the lens of the camera heats up in the sun; according to the expert, the road hardly moves with temperature and the system compensates this movement well. Normal, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-48", "start": "2026-09-07T03:30Z", "end": "2026-09-07T16:00Z", "family": "Visibility", "description": "By day, from sunrise to mid-afternoon: the error_value of all sensors, references included, rises at sunrise and comes back down in the afternoon; the reference R3 is rejected almost all day, as every day (Target issue, separate label). PA_9 is rejected at sunrise with the references PA_8 and R2; at 06:45 the compensation retains no reference and PB4 is rejected, the roadway covered with bright reflections of the low sun on the 06:45 image; PA2_2 and PA3_2 are rejected at 06:00, PB1_2 with the reference PA1_2 at 05:00 and 10:15, when PA2_2 and PA3_2 make a small jump; at 09:45, a vehicle passes in front of R1 and R2 (separate label). Before the outage of 09/09, the reflectors are seen less well by day than after (08/09 thumbnails compared with 23/09) and these rejections are more numerous. The corrected curves of the targets are more agitated than at night, without a clean step; the camera moves in the morning and the compensation follows (separate label, thermal movement of the camera). Deduced cause: daytime heat shimmer above the road, which the expert gives as a trait of the station (curve noise and error_value rising by day only), and the low sun at sunrise; it cannot be seen directly, the images show the scene in full daylight. Overall compensated: the structure is not moving, do not trust the small oscillations of the day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-49", "start": "2026-09-07T04:00Z", "end": "2026-09-07T12:45Z", "family": "Visibility", "description": "The reference PA_8 is rejected for a good part of the morning, intermittently: the compensation runs without it (and without R3), and the basis changes. The targets PA_9, its neighbour in the image, and PA2_2, PA3_2 rise on corr_x and y stay there as long as PA_8 is missing, PA_9 especially, less markedly than on the two following days; they come back in the afternoon when PA_8 is taken again. On the daytime thumbnails from before the outage (08/09 and 09/09 at 09:00), the spot of PA_8 is only a dull point on the light background, while it is sharp on that of 23/09 at the same hour. Cause: PA_8 poorly seen in full daylight. This is not a movement: curves of PA_9, PA2_2 and PA3_2 not to be trusted over the range.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-50", "start": "2026-09-07T06:45Z", "end": "2026-09-07T15:00Z", "family": "Visibility", "description": "PB1_2 alone, by day: its error_value rises much more than that of its neighbours PB2_2 and PB4, and its corr_x moves away on one side from the morning to mid-afternoon, then comes back. As on the other days, the target is at the edge of the pavement kerb, bright in full sun (28/09 daytime thumbnails: its dark spot touches the light band of the kerb); at night, under the infrared, the spot is sharp. Cause deduced from the thumbnails: the daylight on the neighbouring kerb. The structure is not moving: do not trust the curve of PB1_2 over the range.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-51", "start": "2026-09-07T09:30Z", "end": "2026-09-07T10:00Z", "family": "Obstruction", "description": "A vehicle passes in front of the references R1 and R2, seen as translucent on their 09:45 thumbnails, a bus and cars on the image: R1 and R2 are rejected, R3 set aside, the basis changes for the time of one measurement; the targets barely move, PB4 a little on corr_x. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-52", "start": "2026-09-07T14:00Z", "end": "2026-09-07T14:45Z", "family": "Obstruction", "description": "A group of passers-by crosses in front of the targets: on the 14:30 image, a line of translucent silhouettes passes in front of PA3_2, PB_5, PB4 and the reference R2. PA_9 is rejected at 14:15 and 14:30, PA3_2, PB4 and R2 at 14:30; PB_5, which stayed under the threshold, makes a small jump on corr_x. Everything comes back at the next measurement. The structure is not moving: points to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-53", "start": "2026-09-08T02:30Z", "end": "2026-09-08T03:15Z", "family": "Obstruction", "description": "At night, PB1_2 alone: its corr_x makes a jump at 02:45, then its corr_y at 03:00, and everything comes back at 03:15; the other targets and the compensation do not move. The 02:45 thumbnail shows no clear change in its spot. Cause not established, not seen: a passing disturbance in front of the spot, like those that come back at the same hour of the night on the kerb targets (PB1_2 on 03/09 and 01/10, PB2_2 on 24 and 25/09). The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-54", "start": "2026-09-08T03:30Z", "end": "2026-09-08T15:30Z", "family": "Visibility", "description": "By day, from sunrise to mid-afternoon: the error_value of all sensors, references included, rises at sunrise and comes back down in the afternoon; the reference R3 is rejected almost all day, as every day (Target issue, separate label). PA_9 is rejected almost throughout sunrise; at 05:30, PA2_2, PA3_2, PB4 and PB_5 are rejected with R2 and PA_8; at 06:00 and from 06:30 to 06:45, the compensation retains no reference (R1, R2, PA_8, R3 rejected), PA3_2 is rejected at 06:00, PB4 at 06:30; R1, right at its threshold, is rejected several times in the morning; from 08:15 to 09:00, vehicles pass in front of R2 (two separate labels). Before the outage of 09/09, the reflectors are seen less well by day than after (09:00 thumbnails compared with 23/09). The corrected curves of the targets are more agitated than at night, without a clean step; the camera moves in the morning and the compensation follows (separate label, thermal movement of the camera). Deduced cause: daytime heat shimmer above the road, which the expert gives as a trait of the station (curve noise and error_value rising by day only), and the low sun at sunrise; it cannot be seen directly, the images show the scene in full daylight. Overall compensated: the structure is not moving, do not trust the small oscillations of the day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-55", "start": "2026-09-08T03:45Z", "end": "2026-09-08T13:00Z", "family": "Thermal displacement", "description": "At sunrise, the camera nods down: its tilt rises, with its scale and its rotation, and all the raw positions, references included, move together in the image; the tilt peaks in early morning, goes below its night level in late morning, then comes back to it in early afternoon, as every day, while the housing temperature rises; the scale and the pan come back more slowly, until mid-afternoon. The compensation follows: the corrected curves of the targets do not show this wave, only the daytime agitation (separate label). Deduced cause: the mount or the lens of the camera heats up in the sun; according to the expert, the road hardly moves with temperature and the system compensates this movement well. Normal, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-56", "start": "2026-09-08T04:00Z", "end": "2026-09-08T14:30Z", "family": "Visibility", "description": "The reference PA_8 is rejected almost all day: the compensation runs without it (and without R3), and the basis changes. The targets PA_9, its neighbour in the image, and PA2_2, PA3_2 rise on corr_x and y stay there as long as PA_8 is missing, PA_9 especially; they come back in the afternoon when PA_8 is taken again. On the daytime thumbnails from before the outage (08/09 and 09/09 at 09:00), the spot of PA_8 is only a dull point on the light background, while it is sharp on that of 23/09 at the same hour. Cause: PA_8 poorly seen in full daylight. This is not a movement: curves of PA_9, PA2_2 and PA3_2 not to be trusted over the range.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-57", "start": "2026-09-08T06:45Z", "end": "2026-09-08T16:00Z", "family": "Visibility", "description": "PB1_2 alone, by day: its error_value rises much more than that of its neighbours PB2_2 and PB4, and its corr_x moves away on one side from the morning to late afternoon; its corr_y also drops and comes back only in part in the evening. As on the other days, the target is at the edge of the pavement kerb, bright in full sun (28/09 daytime thumbnails: its dark spot touches the light band of the kerb); at night, under the infrared, the spot is sharp. Cause deduced from the thumbnails: the daylight on the neighbouring kerb. The structure is not moving: do not trust the curve of PB1_2 over the range.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-58", "start": "2026-09-08T08:00Z", "end": "2026-09-08T08:45Z", "family": "Obstruction", "description": "Vehicles pass in front of the reference R2: a line of bright reflections near it and PB4 on the 08:15 image, a flash on R2 at 08:30. R2 is rejected two measurements in a row; at 08:15, with R1 right at its threshold and PA_8 poorly seen that day, the compensation retains no reference; at 08:30, PA_9, PA2_2 and PB_5 make a small jump on corr_x. Everything comes back at 08:45. The structure is not moving: points to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-59", "start": "2026-09-08T08:45Z", "end": "2026-09-08T09:15Z", "family": "Obstruction", "description": "A vehicle passes in front of PA3_2 and the reference R2, seen as translucent on the 09:00 image: PA_9 and PA3_2 are rejected for the time of one measurement, with R2, R3 and PA_8; PA2_2 makes a small jump on corr_x. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-60", "start": "2026-09-08T13:15Z", "end": "2026-09-08T13:45Z", "family": "Obstruction", "description": "Passers-by walk on the pavement near PA2_2 and the reference PA1_2, blurry silhouettes on the 13:30 image: PA1_2 is rejected and PA2_2 makes a jump on corr_x for the time of one measurement. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-61", "start": "2026-09-09T03:30Z", "end": "2026-09-09T15:45Z", "family": "Visibility", "description": "By day, from sunrise to mid-afternoon: the error_value of all sensors, references included, rises at sunrise and comes back down in the afternoon; the reference R3 is rejected almost all day, as every day (Target issue, separate label). PA_9 is rejected throughout sunrise; from 05:30 to 06:30, the compensation retains no reference (PA1_2, PA_8, R2, R3 rejected), PA2_2, PA3_2 and PB_5 are rejected at 05:45, PB4 at 05:15, and at 06:30 PA2_2, PA3_2, PB4 and PB_5 make a small jump. Before the outage, the reflectors are seen less well by day than after (09:00 thumbnails compared with 23/09). The corrected curves of the targets are more agitated than at night, without a clean step; the camera moves in the morning and the compensation follows (separate label, thermal movement of the camera). Deduced cause: daytime heat shimmer above the road, which the expert gives as a trait of the station (curve noise and error_value rising by day only), and the low sun at sunrise; it cannot be seen directly, the images show the scene in full daylight. Overall compensated: the structure is not moving, do not trust the small oscillations of the day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-62", "start": "2026-09-09T03:45Z", "end": "2026-09-09T13:30Z", "family": "Thermal displacement", "description": "At sunrise, the camera nods down: its tilt rises, with its scale and its rotation, and all the raw positions, references included, move together in the image; the tilt peaks in early morning, goes below its night level in late morning, then comes back to it in early afternoon, as every day, while the housing temperature rises; the scale and the pan come back more slowly, until mid-afternoon. The compensation follows: the corrected curves of the targets do not show this wave, only the daytime agitation (separate label). Deduced cause: the mount or the lens of the camera heats up in the sun; according to the expert, the road hardly moves with temperature and the system compensates this movement well. Normal, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-63", "start": "2026-09-09T03:45Z", "end": "2026-09-09T15:00Z", "family": "Visibility", "description": "The reference PA_8 is rejected almost all day: the compensation runs without it (and without R3), and the basis changes. The targets PA_9, its neighbour in the image, and PA2_2, PA3_2 rise on corr_x and y stay there as long as PA_8 is missing, PA_9 especially; they come back in the afternoon when PA_8 is taken again. On the daytime thumbnails from before the outage (08/09 and 09/09 at 09:00), the spot of PA_8 is only a dull point on the light background, while it is sharp on that of 23/09 at the same hour. Cause: PA_8 poorly seen in full daylight. This is not a movement: curves of PA_9, PA2_2 and PA3_2 not to be trusted over the range.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-64", "start": "2026-09-09T07:30Z", "end": "2026-09-09T21:45Z", "family": "Visibility", "description": "PB1_2 alone, by day: its error_value rises much more than that of its neighbours PB2_2 and PB4, and its corr_x moves away on one side from mid-morning; in the evening, it comes back only in part, and stays lower until the night outage. As on the other days, the target is at the edge of the pavement kerb, bright in full sun; at night, under the infrared, the spot is sharp. Cause deduced from the thumbnails: the daylight on the neighbouring kerb, without the evening return being complete. The structure is not moving: do not trust the curve of PB1_2 over the range.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-65", "start": "2026-09-09T09:30Z", "end": "2026-09-09T10:00Z", "family": "Obstruction", "description": "A vehicle passes in front of PA3_2, R2, PB4 and R1, seen as translucent on the 09:45 image: PA3_2 and PB4 are rejected, with the references PB3_2, R2 and PA_8, and the compensation retains no reference; PB_5 makes a jump on corr_x and corr_y. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-66", "start": "2026-09-09T10:00Z", "end": "2026-09-09T10:30Z", "family": "Obstruction", "description": "A vehicle passes in front of PB4, PB_5 and R2, seen as translucent on the 10:15 image, with a bright reflection: PA3_2 and PB4 are rejected with R2; PA_9, which stayed under the threshold, locks onto something other than its spot and its corr_x, corr_y make a very large jump; PA2_2 and PB2_2 make a small jump. Everything comes back at the next measurement. The structure is not moving: this point of PA_9 is false, to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-67", "start": "2026-09-09T12:00Z", "end": "2026-09-09T12:30Z", "family": "Obstruction", "description": "A group of passers-by crosses in front of the targets, a line of translucent silhouettes on the 12:15 image, in front of PA3_2, R2, PB_5 and PB4: PA3_2 and PB4 are rejected with R2; the other references move a little, and PB1_2 makes a small jump on corr_x. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-68", "start": "2026-09-09T12:30Z", "end": "2026-09-09T13:00Z", "family": "Obstruction", "description": "A large vehicle (bus or truck) passes in front of PA3_2, R2, PB4 and R1, seen as translucent on the 12:45 image: PA3_2 and PB4 are rejected with R1 and R2, the compensation retains no reference, and PA2_2 and PB_5 make a jump on corr_x. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-69", "start": "2026-09-09T21:45Z", "end": "2026-09-22T08:45Z", "family": "No data", "description": "No measurement for any sensor for nearly two weeks; the metrics and the images stop with the measurements. Just before, the battery empties and the voltage falls well below its level of the other nights (metrics): the station shuts down for lack of energy. It restarts on 22/09 in the morning, the battery full and a voltage much higher than usual, as on an external power supply; on return, the tilt of the camera is no longer at the same night level as before the outage. The curves are missing: the station ran out of energy, to be checked (battery, panel, possible intervention).", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-70", "start": "2026-09-09T21:45Z", "end": "2026-09-22T09:00Z", "family": "Displacement", "description": "During the outage from 09/09 to 22/09, the corrected curves change level and do not resume the one from before: corr_y of PB1_2 markedly lower, corr_y of PB2_2 lower, corr_x of PB_5, PA2_2, PA3_2 and PB4 lower, corr_y of PA_9 higher, from the first measurement. The jump is hidden in the outage; the step lasts until the end of the batch, except for PB_5, back near its level from before on 30/09 (labelled separately), and PB1_2 keeps going down from 25 to 29/09 (labelled separately). The camera is no longer at the same night level (tilt and rotation shifted). Relative to the overall movement of the targets, the raw position of PB1_2 has shifted alone, and the references PA_8 and PA1_2 a little relative to each other, which changes the basis of the other targets. After the restart, the reflectors are seen better, by day and by night, and the error_value of all sensors is lower; the housing temperature has the same cycle before and after. Cause not established: an intervention on the station during the outage (adjustment, cleaning), or, for PB1_2, a movement of the kerb. For the targets that follow the basis, this is not a movement of the structure: do not compare the levels from before and after the outage, to be checked with the operator; for PB1_2, a movement is not excluded: look at the target, keep watching.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-71", "start": "2026-09-22T08:45Z", "end": "2026-09-22T12:30Z", "family": "Thermal displacement", "description": "At the restart, in mid-morning, the camera is still nodded down: its tilt, high, comes back down until noon, goes a little below its evening level then returns to it in early afternoon, with its scale (which comes back more slowly), and all the raw positions, references included, move together in the image, like on the other days at the same hour, while the housing temperature rises. The compensation follows: the corrected curves of the targets do not show this wave, only the daytime agitation (separate label). Deduced cause: the mount or the lens of the camera heats up in the sun; according to the expert, the road hardly moves with temperature and the system compensates this movement well. Normal, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-72", "start": "2026-09-22T08:45Z", "end": "2026-09-22T13:00Z", "family": "Visibility", "description": "PB1_2 alone, by day: its error_value rises much more than that of its neighbours PB2_2 and PB4, and its corr_x moves away on one side on the morning of the restart, then comes back in early afternoon. As on the other days, the target is at the edge of the pavement kerb, bright in full sun (28/09 daytime thumbnails: its dark spot touches the light band of the kerb); at night, under the infrared, the spot is sharp. Cause deduced from the thumbnails: the daylight on the neighbouring kerb. The structure is not moving: do not trust the curve of PB1_2 over the range.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-73", "start": "2026-09-22T08:45Z", "end": "2026-09-22T15:15Z", "family": "Visibility", "description": "By day, from the restart in mid-morning to mid-afternoon: the error_value of all sensors, references included, is high at the restart and comes back down in the afternoon; the reference R3 is rejected almost all day, as every day (Target issue, separate label). At the restart, the first measurement, in mid-morning, is rejected for R1, R2 and R3, PB4 and PA3_2, and PB2_2, PB_5 and PB1_2 make a jump, with nothing in particular on the 08:45 image; everything has come back at the next measurement. The corrected curves of the targets are more agitated than at night, without a clean step; the camera moves in the morning and the compensation follows (separate label, thermal movement of the camera). Deduced cause: daytime heat shimmer above the road, which the expert gives as a trait of the station (curve noise and error_value rising by day only), and the low sun at sunrise; it cannot be seen directly, the images show the scene in full daylight. Overall compensated: the structure is not moving, do not trust the small oscillations of the day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-74", "start": "2026-09-22T17:00Z", "end": "2026-09-22T17:30Z", "family": "Obstruction", "description": "At nightfall, the headlights of a vehicle appear very close to the reference R1 (17:15 thumbnail): the measurement of R1 locks onto a headlight, the compensation rejects it and the basis changes; PB4 makes a jump on corr_x for the time of one measurement. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-75", "start": "2026-09-23T04:00Z", "end": "2026-09-23T12:15Z", "family": "Thermal displacement", "description": "At sunrise, the camera nods down: its tilt rises, with its scale and its rotation, and all the raw positions, references included, move together in the image; the tilt peaks in early morning, goes below its night level in late morning, then comes back to it in early afternoon, as every day, while the housing temperature rises; the scale and the pan come back more slowly, until mid-afternoon. The compensation follows: the corrected curves of the targets do not show this wave, only the daytime agitation (separate label). Deduced cause: the mount or the lens of the camera heats up in the sun; according to the expert, the road hardly moves with temperature and the system compensates this movement well. Normal, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-76", "start": "2026-09-23T04:00Z", "end": "2026-09-23T15:15Z", "family": "Visibility", "description": "By day, from sunrise to mid-afternoon: the error_value of all sensors, references included, rises at sunrise and comes back down in the afternoon; the reference R3 is rejected almost all day, as every day (Target issue, separate label). At sunrise, PA_8 and PA_9 are rejected, and PA1_2, PA2_2, PA_9 and PB1_2 make a small jump on corr_x; PB4 is rejected almost the whole first hour with R2; at 09:30, a vehicle reflection on R1 (separate label). The corrected curves of the targets are more agitated than at night, without a clean step; the camera moves in the morning and the compensation follows (separate label, thermal movement of the camera). Deduced cause: daytime heat shimmer above the road, which the expert gives as a trait of the station (curve noise and error_value rising by day only), and the low sun at sunrise; it cannot be seen directly, the images show the scene in full daylight. Overall compensated: the structure is not moving, do not trust the small oscillations of the day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-77", "start": "2026-09-23T08:30Z", "end": "2026-09-23T09:00Z", "family": "Obstruction", "description": "Bright reflections appear near the reference R2 on the 08:45 image, the reflections of a passing vehicle, deduced from their shape and their duration of one measurement: R1 is rejected, R2 set aside by the compensation, and the basis changes; PB4, PB_5, PB2_2 and PB1_2 jump together on corr_x or corr_y for the time of one measurement. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-78", "start": "2026-09-23T09:15Z", "end": "2026-09-23T09:45Z", "family": "Obstruction", "description": "The reference R1 locks onto a bright reflection, that of a passing vehicle (09:30 thumbnail, absent at 09:15): its position jumps, it is rejected and the basis changes for the time of one measurement; PB4 makes a small jump on corr_x. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-79", "start": "2026-09-23T13:15Z", "end": "2026-09-23T13:45Z", "family": "Obstruction", "description": "A vehicle passes in front of PB4 and the reference R2, seen as translucent on the 13:30 image: PB4 and R2 are rejected for the time of one measurement. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-80", "start": "2026-09-23T13:45Z", "end": "2026-09-23T14:15Z", "family": "Obstruction", "description": "A group of passers-by walks along the pavement, a line of translucent silhouettes on the 14:00 image, in front of PA2_2, PB1_2, PB2_2, PB3_2 and PB_5: PA2_2, PA_9, PB1_2, PB2_2 and PB_5 are rejected for the time of one measurement, with the references PA1_2 and PA_8. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-81", "start": "2026-09-24T02:30Z", "end": "2026-10-01T00:00Z", "family": "Obstruction", "description": "PB2_2 alone: at 02:45 on 24/09, its spot shrinks (02:30 and 02:45 thumbnails) and its corr_x makes a large jump to one side, its corr_y a smaller one to the other; it stays at that level all day. On 25/09 at 02:45, at the same hour, it comes back one step toward its neighbours on corr_x, but not all the way to its level from before, and its corr_y goes lower; at the same moment, the corr_y of its neighbour PB1_2 goes down one step (labelled separately); then corr_x of PB2_2 slides slowly downward until 29/09 and rises back over the day of 30/09, without returning to its level from before 24/09. All this time, its night error_value stays above its level from before and its spot stays smaller (23:00 thumbnails from 24/09 to 30/09, compared with 23/09), while those of PB1_2 and PB4 do not change. Cause partly seen: something covers the spot of PB2_2 (smaller spot), of a nature not seen on the images; the two steps fall at the same hour of the night as the disturbances of PB1_2 on 03/09, 08/09 and 01/10. The structure is not moving: do not trust the curve of PB2_2 over the range; look at what is in front of the target.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-82", "start": "2026-09-24T03:45Z", "end": "2026-09-24T15:15Z", "family": "Visibility", "description": "By day, from sunrise to mid-afternoon: the error_value of all sensors, references included, rises at sunrise and comes back down in the afternoon; the reference R3 is rejected almost all day, as every day (Target issue, separate label). At sunrise, PA_8 and PA_9 are rejected, and several targets make a small jump on corr_x; R2 is rejected almost the whole first hour, PB4 with it at times, and PB1_2 once in mid-morning. The corrected curves of the targets are more agitated than at night, without a clean step; the camera moves in the morning and the compensation follows (separate label, thermal movement of the camera). Deduced cause: daytime heat shimmer above the road, which the expert gives as a trait of the station (curve noise and error_value rising by day only), and the low sun at sunrise; it cannot be seen directly, the images show the scene in full daylight. Overall compensated: the structure is not moving, do not trust the small oscillations of the day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-83", "start": "2026-09-24T04:00Z", "end": "2026-09-24T12:30Z", "family": "Thermal displacement", "description": "At sunrise, the camera nods down: its tilt rises, with its scale and its rotation, and all the raw positions, references included, move together in the image; the tilt peaks in early morning, goes below its night level in late morning, then comes back to it in early afternoon, as every day, while the housing temperature rises; the scale and the pan come back more slowly, until mid-afternoon. The compensation follows: the corrected curves of the targets do not show this wave, only the daytime agitation (separate label), except PB1_2 and PB2_2, at the bottom of the image, whose corr_y rises at the height of the morning movement, following the overall movement, their raw position moving slightly away from it, then comes back in late morning: the compensation follows almost everything (according to the expert, the effect is well compensated); only a small deviation at the bottom of the image remains. Deduced cause: the mount or the lens of the camera heats up in the sun; according to the expert, the road hardly moves with temperature and the system compensates this movement well. Normal, nothing to do.", "review": "2026-11-05"}
{"id": "WILDSTRUBEL-84", "start": "2026-09-24T04:30Z", "end": "2026-09-24T12:15Z", "family": "Visibility", "description": "PB1_2 alone, by day: its error_value rises much more than that of its neighbours PB2_2 and PB4, and its corr_x moves away on one side from sunrise to noon, then comes back. As on the other days, the target is at the edge of the pavement kerb, bright in full sun (28/09 daytime thumbnails: its dark spot touches the light band of the kerb); at night, under the infrared, the spot is sharp. Cause deduced from the thumbnails: the daylight on the neighbouring kerb. The structure is not moving: do not trust the curve of PB1_2 over the range.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-85", "start": "2026-09-24T05:30Z", "end": "2026-09-24T06:00Z", "family": "Visibility", "description": "At sunrise, the shadow of a tree, long and moving with the low sun, covers PB2_2, PB3_2 and PB_5, with patches of sun through the foliage near PB2_2 (05:45 image, compared with those of 05:30 and 06:00): PB2_2, which stayed under the threshold, locks onto something other than its spot and its corr_x, corr_y make a very large jump for the time of one measurement; PA3_2 is rejected, PB4 and R2 too, as every morning at this hour. Everything comes back at the next measurement. Cause seen: the light of the low sun on the fixed scene, not a passing object. The structure is not moving: this point of PB2_2 is false, to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-86", "start": "2026-09-24T09:00Z", "end": "2026-09-24T09:30Z", "family": "Obstruction", "description": "Vehicles pass in front of the targets, translucent trails on the 09:15 image: PA3_2, PB4 and PB1_2 are rejected, with the references R1, R2 and R3; PB_5, which stayed under the threshold, locks onto something other than its spot and its corr_x, corr_y make a very large jump, PB2_2 a small one. Everything comes back at the next measurement. The structure is not moving: this point of PB_5 is false, to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-87", "start": "2026-09-24T09:30Z", "end": "2026-09-24T10:15Z", "family": "Obstruction", "description": "A car with its lights on passes very close to PB4 and the reference R2, seen as translucent on the 09:45 image: PB4 and R2 are rejected; at 10:00, R2 still is, reflections of the car near it (10:00 image). Everything comes back at 10:15. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-88", "start": "2026-09-24T14:00Z", "end": "2026-09-24T14:30Z", "family": "Visibility", "description": "At 14:15, the image is very dark over the whole scene (14:15 image), with no obstacle: PA_9 is rejected for the time of one measurement, with the references PA1_2 and PA_8. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-89", "start": "2026-09-25T02:30Z", "end": "2026-09-29T23:00Z", "family": "Displacement", "description": "PB1_2: after the shift at the restart, its corr_y goes down one step on 25/09 at 02:45, at the same moment as its neighbour PB2_2 rises back in part (labelled separately), then keeps going down slowly, night after night, until the evening of 29/09, and stabilises; with no return until the end of the batch. Its raw position goes down alone relative to the overall movement of the targets; PA_9 and PB4 move much less, and the pose of the camera at night barely moves. Its night error_value stays at its level until 28/09 and rises a little on 29/09; its spot stays sharp at night, a little less round on 29/09 (23:00 thumbnails from 23/09 to 29/09). The housing temperature follows the same cycle every day, with no lasting drop: it does not explain the descent. Cause not established (the target had already been disturbed in early September, at the same hour of the night). A movement of the kerb is not excluded: look at the target PB1_2, keep watching.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-90", "start": "2026-09-25T03:45Z", "end": "2026-09-25T15:15Z", "family": "Visibility", "description": "By day, from sunrise to mid-afternoon: the error_value of all sensors, references included, rises at sunrise and comes back down in the afternoon; the reference R3 is rejected almost all day, as every day (Target issue, separate label). At sunrise, PA_8 and PA_9 are rejected and several targets make a small jump on corr_x; in early morning, PB4, PA3_2, PA_9 and PB2_2 are rejected in turn; at 07:15, a white car on PB2_2 and PB3_2 (separate label). The corrected curves of the targets are more agitated than at night, without a clean step; the camera moves in the morning and the compensation follows (separate label, thermal movement of the camera). Deduced cause: daytime heat shimmer above the road, which the expert gives as a trait of the station (curve noise and error_value rising by day only), and the low sun at sunrise; it cannot be seen directly, the images show the scene in full daylight. Overall compensated: the structure is not moving, do not trust the small oscillations of the day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-91", "start": "2026-09-25T04:15Z", "end": "2026-09-25T12:15Z", "family": "Thermal displacement", "description": "At sunrise, the camera nods down: its tilt rises, with its scale and its rotation, and all the raw positions, references included, move together in the image; the tilt peaks in early morning, goes below its night level in late morning, then comes back to it in early afternoon, as every day, while the housing temperature rises; the scale and the pan come back more slowly, until mid-afternoon. The compensation follows: the corrected curves of the targets do not show this wave, only the daytime agitation (separate label). Deduced cause: the mount or the lens of the camera heats up in the sun; according to the expert, the road hardly moves with temperature and the system compensates this movement well. Normal, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-92", "start": "2026-09-25T07:00Z", "end": "2026-09-25T07:30Z", "family": "Obstruction", "description": "A white car passes on the lane along the pavement, seen on the 07:15 image, over PB2_2 and the reference PB3_2: they are rejected for the time of one measurement, the other targets do not move. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-93", "start": "2026-09-25T09:00Z", "end": "2026-09-25T09:30Z", "family": "Obstruction", "description": "A line of bright points, the lights or reflectors of a passing vehicle, passes at the top left of the scene near PA3_2 (09:15 image): PA3_2 is rejected for the time of one measurement. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-94", "start": "2026-09-25T12:00Z", "end": "2026-09-25T12:30Z", "family": "Obstruction", "description": "A vehicle passes in front of the references R2 and R1, seen as translucent on the 12:15 image: R1 and R2 are rejected and the basis changes; PB4 especially, PB_5, PB2_2, PA3_2 and PB1_2 jump together on corr_x or corr_y for the time of one measurement. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-95", "start": "2026-09-25T13:00Z", "end": "2026-09-25T13:30Z", "family": "Obstruction", "description": "Vehicles pass at the top of the scene near PA_9, translucent on the 13:15 image: PA_9 is rejected for the time of one measurement. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-96", "start": "2026-09-26T03:45Z", "end": "2026-09-26T07:45Z", "family": "Obstruction", "description": "At sunrise, a line of cars parked along the pavement hides the reference PB3_2 and the targets PA2_2, PA3_2, PB2_2 and PB_5, seen on the images from 04:00 to 07:00; a last car stays in front of PA_9 until 07:15. These sensors have their error_value almost at 1 and are rejected almost throughout the range, PA_9 until 07:30. Without PB3_2, and with PA_8 and R2 rejected at sunrise, the compensation runs on few references, then on none at 05:15 (no recalibration): PB1_2 then makes a large jump, PB4 and PA1_2 smaller ones. Everything comes back at 07:45. The structure is not moving: ignore the curves of these targets over the range.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-97", "start": "2026-09-26T03:45Z", "end": "2026-09-26T15:15Z", "family": "Visibility", "description": "By day, from sunrise to mid-afternoon: the error_value of all sensors, references included, rises at sunrise and comes back down in the afternoon; the reference R3 is rejected almost all day, as every day (Target issue, separate label). The morning is mixed with the parked cars in front of some of the targets (separate label); R2 is rejected almost the whole first hour. The corrected curves of the targets are more agitated than at night, without a clean step; the camera moves in the morning and the compensation follows (separate label, thermal movement of the camera). Deduced cause: daytime heat shimmer above the road, which the expert gives as a trait of the station (curve noise and error_value rising by day only), and the low sun at sunrise; it cannot be seen directly, the images show the scene in full daylight. Overall compensated: the structure is not moving, do not trust the small oscillations of the day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-98", "start": "2026-09-26T04:15Z", "end": "2026-09-26T12:30Z", "family": "Thermal displacement", "description": "At sunrise, the camera nods down: its tilt rises, with its scale and its rotation, and all the raw positions, references included, move together in the image; the tilt peaks in early morning, goes below its night level in late morning, then comes back to it in early afternoon, as every day, while the housing temperature rises; the scale and the pan come back more slowly, until mid-afternoon. The compensation follows: the corrected curves of the targets do not show this wave, only the daytime agitation (separate label). Deduced cause: the mount or the lens of the camera heats up in the sun; according to the expert, the road hardly moves with temperature and the system compensates this movement well. Normal, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-99", "start": "2026-09-26T07:30Z", "end": "2026-09-26T08:00Z", "family": "Obstruction", "description": "Bright reflections appear near the reference R2 and PB4 (07:45 image), those of a passing vehicle, absent at the measurement before and at the one after: PB4 and R2 are rejected for the time of one measurement. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-100", "start": "2026-09-26T10:30Z", "end": "2026-09-26T11:00Z", "family": "Obstruction", "description": "A vehicle with its lights on passes in front of PB4 and the reference R2, seen as translucent on the 10:45 image: PB4 and R2 are rejected for the time of one measurement. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-101", "start": "2026-09-27T04:00Z", "end": "2026-09-27T15:15Z", "family": "Visibility", "description": "By day, from sunrise to mid-afternoon: the error_value of all sensors, references included, rises at sunrise and comes back down in the afternoon; the reference R3 is rejected almost all day, as every day (Target issue, separate label). At sunrise, PA_8 and PA_9 are rejected, and PA1_2, PA2_2, PA3_2 and PB_5 make a small jump on corr_x, PA2_2 also on corr_y; PB2_2, PB4 and PA3_2 are rejected once, PB4 then almost the whole first hour with R2. The corrected curves of the targets are more agitated than at night, without a clean step; the camera moves in the morning and the compensation follows (separate label, thermal movement of the camera). Deduced cause: daytime heat shimmer above the road, which the expert gives as a trait of the station (curve noise and error_value rising by day only), and the low sun at sunrise; it cannot be seen directly, the images show the scene in full daylight. Overall compensated: the structure is not moving, do not trust the small oscillations of the day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-102", "start": "2026-09-27T04:15Z", "end": "2026-09-27T12:30Z", "family": "Thermal displacement", "description": "At sunrise, the camera nods down: its tilt rises, with its scale and its rotation, and all the raw positions, references included, move together in the image; the tilt peaks in early morning, goes below its night level in late morning, then comes back to it in early afternoon, as every day, while the housing temperature rises; the scale and the pan come back more slowly, until mid-afternoon. The compensation follows: the corrected curves of the targets do not show this wave, only the daytime agitation (separate label), except PB1_2 and PB2_2, at the bottom of the image, whose corr_y rises at the height of the morning movement, following the overall movement, their raw position moving slightly away from it, then comes back in late morning: the compensation follows almost everything (according to the expert, the effect is well compensated); only a small deviation at the bottom of the image remains. Deduced cause: the mount or the lens of the camera heats up in the sun; according to the expert, the road hardly moves with temperature and the system compensates this movement well. Normal, nothing to do.", "review": "2026-11-05"}
{"id": "WILDSTRUBEL-103", "start": "2026-09-27T06:45Z", "end": "2026-09-27T07:15Z", "family": "Obstruction", "description": "Bright reflections line up on the central reservation near the reference R2 (07:00 image), the reflections of passing vehicles: R1, R2 and R3 are rejected, PB4 and PA3_2 too, and the basis changes; PB2_2 and PB_5 jump on corr_x and corr_y for the time of one measurement. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-104", "start": "2026-09-27T08:00Z", "end": "2026-09-27T08:30Z", "family": "Obstruction", "description": "A white van stops along the pavement, seen on the 08:15 image, in front of PA2_2, PA3_2 and the reference PA1_2, its lights on PA3_2: these three sensors are rejected for the time of one measurement. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-105", "start": "2026-09-28T04:00Z", "end": "2026-09-28T15:15Z", "family": "Visibility", "description": "By day, from sunrise to mid-afternoon: the error_value of all sensors, references included, rises at sunrise and comes back down in the afternoon; the reference R3 is rejected almost all day, as every day (Target issue, separate label). At sunrise, PA_8 and PA_9 are rejected, PB4 and PA2_2 make a jump; PB4 is rejected almost the whole first hour with R2; at 05:30, then from 06:00 to 06:15 and at 08:15, vehicles pass in front of the references in the middle of the roadway (three separate labels). The corrected curves of the targets are more agitated than at night, without a clean step; the camera moves in the morning and the compensation follows (separate label, thermal movement of the camera). Deduced cause: daytime heat shimmer above the road, which the expert gives as a trait of the station (curve noise and error_value rising by day only), and the low sun at sunrise; it cannot be seen directly, the images show the scene in full daylight. Overall compensated: the structure is not moving, do not trust the small oscillations of the day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-106", "start": "2026-09-28T04:15Z", "end": "2026-09-28T12:30Z", "family": "Thermal displacement", "description": "At sunrise, the camera nods down: its tilt rises, with its scale and its rotation, and all the raw positions, references included, move together in the image; the tilt peaks in early morning, goes below its night level in late morning, then comes back to it in early afternoon, as every day, while the housing temperature rises; the scale and the pan come back more slowly, until mid-afternoon. The compensation follows: the corrected curves of the targets do not show this wave, only the daytime agitation (separate label), except PB1_2 and PB2_2, at the bottom of the image, whose corr_y rises at the height of the morning movement, following the overall movement, their raw position moving slightly away from it, then comes back in late morning: the compensation follows almost everything (according to the expert, the effect is well compensated); only a small deviation at the bottom of the image remains. Deduced cause: the mount or the lens of the camera heats up in the sun; according to the expert, the road hardly moves with temperature and the system compensates this movement well. Normal, nothing to do.", "review": "2026-11-05"}
{"id": "WILDSTRUBEL-107", "start": "2026-09-28T05:15Z", "end": "2026-09-28T05:45Z", "family": "Obstruction", "description": "At sunrise, R1 and R2 are rejected together for the time of one measurement and the basis changes: PB_5 especially, PB1_2, PB2_2 and PA3_2 jump on corr_x and corr_y. On the 05:30 image, translucent cars at the top of the scene and a blurry dark shape at the bottom; deduced cause: the signature of the passages seen from the station (R1 and R2, on the same lane, lost together, back at the next measurement). PB4 and R2 are rejected before and after, as every morning. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-108", "start": "2026-09-28T05:45Z", "end": "2026-09-28T06:30Z", "family": "Obstruction", "description": "At sunrise, vehicles pass: on the 06:00 image, cars and a bus at the top of the scene, and lines of bright points, their translucent reflections, on the roadway around PA3_2 and the reference R2, absent at 05:45; at 06:15, other reflections near R2 and PB4. PB3_2, PB1_2, PB2_2, PB_5 and PA3_2 are rejected at 06:00 with R2, and R1 is at 06:15; PB4 and R2, rejected at every measurement of this first hour, are so as every morning. Everything comes back at 06:30. Cause seen: the reflections of the passing vehicles (a reflection carried by a vehicle is the passage), not the sun on the roadway, which does not change from one measurement to the next. The structure is not moving: points to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-109", "start": "2026-09-28T07:15Z", "end": "2026-09-28T07:45Z", "family": "Obstruction", "description": "Bright reflections appear near the reference R2 and PB4 (07:30 image), those of a passing vehicle, absent at the measurement before and at the one after: PB4 and R2 are rejected for the time of one measurement. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-110", "start": "2026-09-28T08:00Z", "end": "2026-09-28T08:30Z", "family": "Obstruction", "description": "A line of bright reflections, those of a passing vehicle, crosses the roadway near PA3_2 and the reference R2 (08:15 image): PA3_2 is rejected and R2 set aside by the compensation for the time of one measurement; the other targets barely move. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-111", "start": "2026-09-29T02:30Z", "end": "2026-09-29T05:15Z", "family": "No data", "description": "No measurement for any sensor at the end of the night, until sunrise. The station shuts down for lack of energy: just before the outage, the battery is at its lowest and the voltage below its usual night level, and the metrics stop with the measurements; it restarts at sunrise with the solar charge. The curves are missing; nothing moves on the structure, but the station does not have enough energy to last the night: to be checked (battery, panel).", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-112", "start": "2026-09-29T05:15Z", "end": "2026-09-29T09:15Z", "family": "Visibility", "description": "PB1_2 alone, by day: its error_value rises much more than that of its neighbours PB2_2 and PB4, and its corr_x moves away on one side from sunrise to mid-morning, then comes back. As on the other days, the target is at the edge of the pavement kerb, bright in full sun (28/09 daytime thumbnails: its dark spot touches the light band of the kerb); at night, under the infrared, the spot is sharp. Cause deduced from the thumbnails: the daylight on the neighbouring kerb. The structure is not moving: do not trust the curve of PB1_2 over the range.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-113", "start": "2026-09-29T05:15Z", "end": "2026-09-29T12:30Z", "family": "Thermal displacement", "description": "At the restart at sunrise, after the night outage, the camera nods down: its tilt rises, with its scale and its rotation, and all the raw positions, references included, move together in the image; the tilt peaks in early morning, goes below its night level in late morning, then comes back to it in early afternoon, as every day, while the housing temperature rises; the scale and the pan come back more slowly, until mid-afternoon. The compensation follows: the corrected curves of the targets do not show this wave, only the daytime agitation (separate label). Deduced cause: the mount or the lens of the camera heats up in the sun; according to the expert, the road hardly moves with temperature and the system compensates this movement well. Normal, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-114", "start": "2026-09-29T05:15Z", "end": "2026-09-29T15:15Z", "family": "Visibility", "description": "By day, from sunrise to mid-afternoon: the error_value of all sensors, references included, rises at sunrise and comes back down in the afternoon; the reference R3 is rejected almost all day, as every day (Target issue, separate label). At the restart after the night outage, PA3_2, PB2_2 and PB4 are rejected, then PB4 at times with R2. The corrected curves of the targets are more agitated than at night, without a clean step; the camera moves in the morning and the compensation follows (separate label, thermal movement of the camera). Deduced cause: daytime heat shimmer above the road, which the expert gives as a trait of the station (curve noise and error_value rising by day only), and the low sun at sunrise; it cannot be seen directly, the images show the scene in full daylight. Overall compensated: the structure is not moving, do not trust the small oscillations of the day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-115", "start": "2026-09-29T10:00Z", "end": "2026-09-29T10:30Z", "family": "Obstruction", "description": "A vehicle passes in front of PB4, the reference R2 and R1, seen as translucent on the 10:15 image: PB4 and PA3_2 are rejected with R2, and PB1_2 makes a small jump on corr_x. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-116", "start": "2026-09-29T10:45Z", "end": "2026-09-29T11:15Z", "family": "Obstruction", "description": "Blurry shapes pass near the reference R2, translucent on the 11:00 image, a passing vehicle: R1, R2 and R3 are rejected and the basis changes; PB4, PB_5, PA3_2, PB1_2 and PB2_2 jump together on corr_x and corr_y for the time of one measurement. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-117", "start": "2026-09-29T16:15Z", "end": "2026-09-29T16:45Z", "family": "Obstruction", "description": "At nightfall, the spot of PA_9 disappears for the time of one measurement (16:30 thumbnail) and that of the reference PA_8 shrinks: PA_9 and PA_8 are rejected, PA2_2 makes a small jump on corr_y. Everything comes back at the next measurement. Cause deduced from the signature of a passage (two neighbouring sensors lost for a single measurement, immediate return), like the passages seen from the station; the night image does not show the object. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-118", "start": "2026-09-29T21:00Z", "end": "2026-09-29T21:30Z", "family": "Obstruction", "description": "At night, the headlights of a vehicle appear very close to PA_9 (21:15 thumbnail, absent at 21:00 and 21:30): its measurement locks onto a headlight while staying under the threshold, and its corr_x, corr_y make a very large jump for the time of one measurement. The other targets and the compensation do not move. The structure is not moving: this point of PA_9 is false, to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-119", "start": "2026-09-30T02:15Z", "end": "2026-09-30T05:45Z", "family": "No data", "description": "No measurement for any sensor at the end of the night, except two isolated measurements at 04:30 and 05:15 before the restart. The station shuts down for lack of energy: just before the outage, the battery is at its lowest and the voltage below its usual night level, and the metrics stop with the measurements; it restarts at sunrise with the solar charge. The curves are missing; nothing moves on the structure, but the station does not have enough energy to last the night: to be checked (battery, panel).", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-120", "start": "2026-09-30T02:15Z", "end": "2026-09-30T16:15Z", "family": "Displacement", "description": "On 30/09, during the night outage and the day that follows, two targets change level, then stay there until the end of the batch: corr_x of PB_5 rises, already higher at the isolated measurement of 04:30 and in place from early morning, near its level from before the outage of 09/09 (this is the return of PB_5 from that shift, labelled separately); corr_x of PA3_2 goes down over the day and stabilises in late afternoon. PB2_2 rises back the same day (separate label); the other targets do not move. Their spots stay sharp and round at night (23:00 thumbnails of 29 and 30/09), their night error_value barely rises, like that of PB4. The thumbnails of that day are darker than those of the previous days. Cause not established (an adjustment of the image capture after the outage, or a movement of these targets). To be checked: look at PB_5 and PA3_2, keep watching the following days.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-121", "start": "2026-09-30T04:30Z", "end": "2026-09-30T12:15Z", "family": "Thermal displacement", "description": "At the restart at sunrise, after the night outage, the camera nods down: its tilt rises, with its scale and its rotation, and all the raw positions, references included, move together in the image; the tilt peaks in early morning, goes below its night level in late morning, then comes back to it in early afternoon, as every day, while the housing temperature rises; the scale and the pan come back more slowly, until mid-afternoon. The compensation follows: the corrected curves of the targets do not show this wave, only the daytime agitation (separate label), except PB1_2 and PB2_2, at the bottom of the image, whose corr_y rises at the height of the morning movement, following the overall movement, their raw position moving slightly away from it, then comes back in late morning: the compensation follows almost everything (according to the expert, the effect is well compensated); only a small deviation at the bottom of the image remains. Deduced cause: the mount or the lens of the camera heats up in the sun; according to the expert, the road hardly moves with temperature and the system compensates this movement well. Normal, nothing to do.", "review": "2026-11-05"}
{"id": "WILDSTRUBEL-122", "start": "2026-09-30T04:30Z", "end": "2026-09-30T15:30Z", "family": "Visibility", "description": "By day, from sunrise to mid-afternoon: the error_value of all sensors, references included, rises at sunrise and comes back down in the afternoon; the reference R3 is rejected almost all day, as every day (Target issue, separate label). At the restart after the night outage, PB1_2 and PB4 are rejected in early morning, PB4 twice; at 13:15 and at 13:45, passages in front of PA1_2 (separate labels); the daytime thumbnails are darker than those of the previous days. The corrected curves of the targets are more agitated than at night, without a clean step; the camera moves in the morning and the compensation follows (separate label, thermal movement of the camera). Deduced cause: daytime heat shimmer above the road, which the expert gives as a trait of the station (curve noise and error_value rising by day only), and the low sun at sunrise; it cannot be seen directly, the images show the scene in full daylight. Overall compensated: the structure is not moving, do not trust the small oscillations of the day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-123", "start": "2026-09-30T07:30Z", "end": "2026-09-30T08:15Z", "family": "Obstruction", "description": "A large vehicle passes in front of PA3_2, the reference R2, PB4 and R1, seen as translucent on the 07:45 and 08:00 images: R1, R2 and R3 are rejected two measurements in a row and the basis changes; PB4 and PB_5 especially, PA3_2, PA2_2, PB2_2 and PB1_2 jump together on corr_x and corr_y. Everything comes back at 08:15. The structure is not moving: points to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-124", "start": "2026-09-30T08:15Z", "end": "2026-09-30T08:45Z", "family": "Obstruction", "description": "A vehicle passes in front of PA3_2, R2 and PB4, seen as translucent on the 08:30 image: PB4 is rejected for the time of one measurement. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-125", "start": "2026-09-30T09:15Z", "end": "2026-09-30T09:45Z", "family": "Obstruction", "description": "Bright reflections of vehicles appear near the references R2 and R1 (09:30 image): R1, R2 and R3 are rejected and the basis changes; PB4, PB_5, PB2_2, PA3_2 and PB1_2 jump together on corr_x and corr_y for the time of one measurement. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-126", "start": "2026-09-30T09:45Z", "end": "2026-09-30T10:15Z", "family": "Obstruction", "description": "A blurry shape passes on the pavement in front of PB2_2 and PB1_2 (10:00 image): they are rejected for the time of one measurement. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-127", "start": "2026-09-30T10:45Z", "end": "2026-09-30T11:30Z", "family": "Obstruction", "description": "For two measurements in a row, R1 and R2 are rejected, with R3, and the basis changes: PB4 and PB_5 especially, PA3_2, PB2_2, PA2_2 and PB1_2 jump together on corr_x and corr_y, then everything comes back at 11:30. The 11:00 and 11:15 images show nothing clear near R1 and R2; cause deduced from the signature of the passages seen from the station (R1 and R2, on the same lane, lost together, back at the next measurement): vehicles passing in front of them. The structure is not moving: points to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-128", "start": "2026-09-30T13:00Z", "end": "2026-09-30T13:30Z", "family": "Obstruction", "description": "A light, blurry shape passes on the pavement to the left of the scene (13:15 image), a passer-by: PA2_2 and PB1_2 are rejected for the time of one measurement, with the references PA1_2 and PA_8. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-129", "start": "2026-09-30T13:30Z", "end": "2026-09-30T14:00Z", "family": "Obstruction", "description": "The reference PA1_2 is rejected alone for the time of one measurement, and the basis changes: PA2_2, its neighbour, makes a small jump on corr_x, PA_9 a smaller one. Nothing clear near it on the 13:45 image, which is darker; cause deduced from the signature of a passage on the pavement (one measurement, immediate return), like the one at 13:15 in front of the same reference. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-130", "start": "2026-10-01T00:00Z", "end": "2026-11-01T00:00Z", "family": "Target issue", "description": "The reference R3 is rejected every measured day, from the first measurement of the morning (04:30 to 05:45) until 13:30-16:30, and almost never at night: most of the measurements from 05:00 to 15:00, often all of them, from 01/10 to 09/10, and at the only daytime measurement of 15/10 (08:00); almost none at night. Its median error_value is markedly above the threshold by day, and a little below it at night. It is as much so on the days when the housing heats up least (on 08/10, almost all its daytime measurements): it is not the daytime sun, it is the target. On its daytime thumbnails, its spot cannot be seen (30/09 at 10:00), while those of R1 and R2 can. The compensation therefore runs every day on five references, and on three or fewer when two others are lost (passages, PA_8 and PA1_2 poorly seen in the sun, separate labels); its entries and exits do not make the targets jump (median jump of the corr no more than a little above the noise, like the other steps of the day, apart from the moments already labelled). No daytime measurement after 15/10: no return seen, until the end of the batch. Cause not established: a reflector too weak for daytime (small, dirty or poorly oriented). Ignore the curve of R3 by day; act on the target (clean it, reorient it or replace it).", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-131", "start": "2026-10-01T02:30Z", "end": "2026-10-01T03:15Z", "family": "Obstruction", "description": "At night, PB1_2 alone: at 02:45 its spot shrinks markedly (02:45 thumbnail, compared with those of 02:30 and 03:00, while those of PB2_2 and PB4 stay whole); its measurement is rejected, then at 03:00 its corr_x makes a spike and it has not come back to its level at 03:15, the last measurement before the night outage. PB2_2, its neighbour, drops on corr_x at the same moment and stays lower until the outage. Cause: something partly covers the spot of PB1_2, of a nature not seen (at the same hour as the obstruction of PB1_2 on 03/09). The structure is not moving: do not trust the curves of PB1_2 and PB2_2 over the range.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-132", "start": "2026-10-01T03:15Z", "end": "2026-10-01T05:45Z", "family": "No data", "description": "No measurement for any sensor during the night, except two isolated measurements at 04:45 and 05:15 before the restart. The station shuts down for lack of energy: just before the outage, the battery is at its lowest and the voltage below its usual night level, and the metrics stop with the measurements; it restarts at sunrise with the solar charge, the compute board still cold (metrics). The curves are missing; nothing moves on the structure, but the station does not have enough energy to last the night: to be checked (battery, panel).", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-133", "start": "2026-10-01T04:45Z", "end": "2026-10-01T12:30Z", "family": "Thermal displacement", "description": "From sunrise until around noon, the camera moves through thermal effect, as every day: the tilt started to rise (the camera nods down) during the end-of-night outage: a little above its night level at the isolated measurement of 04:45, markedly at 06:00; it goes markedly below its level in late morning (at 10:45) and regains at 12:30 its afternoon level, a little below that of the night. The pan drops a little at sunrise (around 05:00-06:00); the rotation goes above its level around 10:00-10:30 and comes back around 12:30; at 08:30, when all the references are lost together, the pose makes a one-measurement jump (separate label); the scale rises with the temperature (markedly at 11:45) and comes back down only in the afternoon, still above its night level at 19:00. In the image, targets and references move together, as a block: y markedly positive at 06:00, then a little negative at 11:00 (ensemble.py --rigide; the targets move away from the block only barely). The housing temperature rises strongly from 04h to 07h (a little less hot at 09h, again at 11h) and stays hot at 12h: the tilt rises with the warming of the morning and comes back while the housing is still hot; the scale follows the temperature and comes back down with it. The compensation follows: on the corrected curves, the tilt wave explains little: before the loss of PA_8, barely more than the noise on each target, except PB1_2 (markedly on corr_x, over few measurements only); what remains is the daytime agitation, the steps of the loss of PA_8 and the common shift taken that day (separate labels). Cause given by the expert for this station (label of 27/09): the camera moves through thermal effect, on its lens or its mount; the road does not move with temperature. Normal, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-134", "start": "2026-10-01T04:45Z", "end": "2026-10-01T15:30Z", "family": "Visibility", "description": "By day, from sunrise to mid-afternoon: the error_value of all sensors, references included, rises at sunrise and comes back down in the afternoon; the reference R3 is rejected almost all day, as every day (Target issue, separate label). The corrected curves of the targets are markedly more agitated than in the evening, without a clean step. At sunrise, PA3_2, PB2_2, PB4 and PB_5 are rejected several times, PB2_2 almost the whole first hour, and the references PA_8 and R2 too; at 14:00, PA_8 is set aside again, PA_9 makes a one-measurement jump. Cause: daytime heat shimmer above the road and the low sun at sunrise, as every day on this station; not seen directly, the images show the scene in full daylight, the spots of the reflectors drowned in the light background. Overall compensated: the structure is not moving, do not trust the small oscillations of the day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-135", "start": "2026-10-01T08:15Z", "end": "2026-10-01T13:15Z", "family": "Visibility", "description": "The reference PA_8 is rejected continuously from 08:30 to 13:00: the compensation runs without it, and the basis changes. PA_9, its neighbour in the image, rises markedly on corr_x and y stays on a step throughout the range; PA2_2, PA3_2, PB2_2 and PB_5 rise a little. At 08:30 and 09:15, all the references are lost together (08:30 image darker, the scene in shadow, with no obstacle): all targets jump for the time of one measurement, PA_9 especially. Everything comes back in part at 13:15, when PA_8 is taken again; PA_9 stays a little above its morning level until the evening. Cause: PA_8 poorly seen in full daylight, its spot faint on the background (14:00 thumbnail). The step is not a movement: curves of PA_9, PA2_2, PA3_2, PB2_2 and PB_5 not to be trusted over the range.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-136", "start": "2026-10-01T13:00Z", "end": "2026-10-01T13:30Z", "family": "Displacement", "description": "When the reference PA_8 is taken again, on 01/10 at 13:15, all the targets find themselves shifted from their level of late September, with the six references in the compensation: corr_y lower for all of them, together; corr_x higher for PA_9 and PA2_2, lower for PB1_2 and PB_5. The shift occurred between the night and noon, under the thermal movement of the morning and the loss of PA_8 (labelled separately): it cannot be dated more precisely. Relative to the overall movement of the targets, the references PA1_2 and PA_8 moved apart a little, mostly vertically, and R1 horizontally, whereas they followed it the evening before: the basis of references has changed a little, and all the targets inherit it together. The shift remains on 02/10 and 03/10 and fades on 04/10 (return labelled separately). Cause not established (a remainder of the camera movement poorly compensated, or a change in the appearance of these references). This is not an established movement of the road: do not read these days as a movement, to be checked with the following days.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-137", "start": "2026-10-01T21:30Z", "end": "2026-10-02T05:15Z", "family": "No data", "description": "No measurement for any sensor during the night, except one isolated measurement at 04:45. The station shuts down for lack of energy: just before the outage, the battery is at its lowest and the voltage below its usual night level, and the metrics stop with the measurements; it restarts at sunrise with the solar charge, the compute board still cold (metrics). The curves are missing; nothing moves on the structure, but the station does not have enough energy to last the night: to be checked (battery, panel).", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-138", "start": "2026-10-02T04:45Z", "end": "2026-10-02T13:00Z", "family": "Thermal displacement", "description": "From sunrise until around noon, the camera moves through thermal effect, as every day: its tilt (the camera nods down) is already high at the isolated measurement of 04:45, highest at 06:15, markedly above its afternoon rest; it goes markedly below this rest in late morning (at 11:30) and regains it at 13:00. The pan is a little lower than in the afternoon in the morning (around 05:15); the rotation goes markedly above its level (at 10:15) and comes back at 13:00; the scale rises with the temperature (markedly at 08:30) and comes back down in the afternoon, at its level around 17:30. In the image, targets and references move together, as a block: y markedly positive at 06:15, then a little negative at 11:30 (ensemble.py --rigide; the targets move away from the block only barely). The housing temperature rises strongly from 04h to 09h and stays hot at 13h: the tilt rises with the warming of the morning and comes back while the housing is still hot; the scale follows the temperature and comes back down with it. The compensation follows: on the corrected curves, the tilt wave explains little, apart from the steps of the loss of PA_8 (from 06:00 to 13:00) and of the first measurements without PA1_2 or PA_8, which make the targets jump; these episodes and the daytime agitation are labelled separately. Cause given by the expert for this station (label of 27/09): the camera moves through thermal effect, on its lens or its mount; the road does not move with temperature. Normal, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-139", "start": "2026-10-02T04:45Z", "end": "2026-10-02T15:30Z", "family": "Visibility", "description": "By day, from sunrise to mid-afternoon: the error_value of all sensors, references included, rises at sunrise and comes back down in the afternoon; the reference R3 is rejected almost all day, as every day (Target issue, separate label). The corrected curves of the targets are markedly more agitated than in the evening, without a clean step. At sunrise, PA3_2 and PA_9 are rejected with the references PA1_2 and PA_8, then PB4 at 06:00 and PB1_2 at 11:30. Cause: daytime heat shimmer above the road and the low sun at sunrise, as every day on this station; not seen directly, the images show the scene in full daylight, the spots of the reflectors drowned in the light background. Overall compensated: the structure is not moving, do not trust the small oscillations of the day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-140", "start": "2026-10-02T06:00Z", "end": "2026-10-02T13:00Z", "family": "Visibility", "description": "The reference PA_8 is set aside at times at sunrise, then continuously from 08:00 to 12:45: the basis changes. PA_9 rises on corr_x, PB1_2 too, PA2_2 a little, in steps that follow the entries and exits of PA_8; at 12:45, all the references are lost together, PA3_2 and PB4 rejected, with no obstacle visible on the image. Everything comes back when PA_8 is taken again at 13:00. Cause: PA_8 poorly seen in full daylight, like the day before. The steps are not a movement: curves of PA_9, PB1_2 and PA2_2 not to be trusted over the range.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-141", "start": "2026-10-02T12:45Z", "end": "2026-10-02T20:15Z", "family": "Obstruction", "description": "The reference PA1_2 is rejected without interruption from 13:00 until the evening outage: its spot is markedly smaller than at the same hours on the other days (18:00 thumbnail, compared with 01/10 and 06/10), while those of PA_8 and PA2_2 are whole. Without it, the basis changes: PA2_2, PA3_2 and PB_5 stay a little lower on corr_x than the evening before, PA_9 too, until the evening, with no return before the outage; the common shift of corr_y of all the targets, for its part, was already there the evening before with all the references (separate label); the next morning, PA1_2 is back. Cause: something partly covers the spot of PA1_2, of a nature not seen. The structure is not moving: slight shift of the curves not to be trusted; look at what is in front of PA1_2 if it comes back.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-142", "start": "2026-10-02T16:15Z", "end": "2026-10-02T16:45Z", "family": "Obstruction", "description": "The headlights of a vehicle appear just above PB4 and PA3_2 (16:30 thumbnails, absent at 16:15): PA3_2 is rejected, PB4, which stayed under the threshold, locks onto a headlight and its corr_x, corr_y make a huge jump for the time of one measurement. The other targets do not move; everything comes back at the next measurement. The structure is not moving: this point of PB4 is false, to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-143", "start": "2026-10-02T20:15Z", "end": "2026-10-03T05:00Z", "family": "No data", "description": "No measurement for any sensor during the night. The station shuts down for lack of energy: just before the outage, the battery is at its lowest and the voltage below its usual night level, and the metrics stop with the measurements; it restarts at sunrise with the solar charge, the compute board still cold (metrics). The curves are missing; nothing moves on the structure, but the station does not have enough energy to last the night: to be checked (battery, panel).", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-144", "start": "2026-10-03T05:00Z", "end": "2026-10-03T07:45Z", "family": "Obstruction", "description": "At sunrise, a line of cars parked along the pavement hides the references PA1_2 and PB3_2 and the targets PA2_2, PA3_2, PB2_2, PB4 and PB_5 (05:30 image, low sun in the field): their measurements are rejected until 06:30; with the other references poorly seen at sunrise, the compensation retains none at several moments (no recalibration at 05:00 and 05:45), and only PB1_2 keeps a curve, which jumps. At 06:45 the cars have left, except one in the background, in front of PA_9 (06:45 image), rejected until 07:30. Everything comes back afterwards. The structure is not moving: ignore the curves over the range.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-145", "start": "2026-10-03T05:00Z", "end": "2026-10-03T12:30Z", "family": "Thermal displacement", "description": "From sunrise until around noon, the camera moves through thermal effect, as every day: its tilt (the camera nods down) is already high at the first recalibrated measurement of 05:20, highest at 06:00, markedly above its afternoon rest; it goes markedly below this rest in late morning (at 10:15) and regains it at 12:30. The pan rises a little in mid-morning (at 10:30) and the rotation markedly (at 10:15), then they come back at 12:30; the scale rises with the temperature (markedly at 09:15) and comes back down in the afternoon. In the image, targets and references move together, as a block: y a little positive at 07:45, after the parked cars leave (separate label), then a little more negative at 10:30 (ensemble.py --rigide; the targets move away from the block only barely). The housing temperature rises strongly from 04h to 09h and stays hot at 12h: the tilt rises with the warming of the morning and comes back while the housing is still hot; the scale follows the temperature and comes back down with it. The compensation follows: on the corrected curves, the tilt wave weighs only a little on each target, on the measurements where R1, R2 and PB3_2 are retained; what remains is the daytime agitation, the parked cars of the morning, the passages and the steps of the loss of PA_8 (separate labels). Cause given by the expert for this station (label of 27/09): the camera moves through thermal effect, on its lens or its mount; the road does not move with temperature. Normal, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-146", "start": "2026-10-03T05:00Z", "end": "2026-10-03T15:15Z", "family": "Visibility", "description": "By day, from sunrise to mid-afternoon: the error_value of all sensors, references included, rises at sunrise and comes back down in the afternoon; the reference R3 is rejected almost all day, as every day (Target issue, separate label). The corrected curves of the targets are markedly more agitated than in the evening, without a clean step. The one-measurement rejections at 08:45, 11:00 and 12:00 are passages (separate labels). Cause: daytime heat shimmer above the road and the low sun at sunrise, as every day on this station; not seen directly, the images show the scene in full daylight, the spots of the reflectors drowned in the light background. Overall compensated: the structure is not moving, do not trust the small oscillations of the day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-147", "start": "2026-10-03T07:30Z", "end": "2026-10-03T08:00Z", "family": "Obstruction", "description": "At 07:45, vehicle reflections appear near the references R1 and R2 (07:45 thumbnails, absent at 07:30): they are set aside with R3, and the basis changes for the time of one measurement; all the targets jump together on corr_x and corr_y, PB4 and PA_9 the most, then come back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-148", "start": "2026-10-03T08:00Z", "end": "2026-10-03T13:45Z", "family": "Visibility", "description": "The reference PA_8 is set aside for most of late morning: the basis changes, and PA_9 rises on corr_x, PB1_2 and PA3_2 a little, in steps that follow the entries and exits of PA_8; everything comes back when it is taken again. Cause: PA_8 poorly seen in full daylight, like the two previous days. The steps are not a movement: curves of PA_9 and PB1_2 not to be trusted over the range.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-149", "start": "2026-10-03T08:30Z", "end": "2026-10-03T09:00Z", "family": "Obstruction", "description": "For the time of one measurement, PB1_2 and PB2_2, at the edge of the pavement, are rejected together; the references and the other targets do not move. On the 08:45 thumbnail, blurry translucent traces pass over PB2_2: probably a passer-by, probable cause, barely visible. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-150", "start": "2026-10-03T10:00Z", "end": "2026-10-03T10:30Z", "family": "Obstruction", "description": "A vehicle passes on the road in front of PA3_2 and PB4 (bright points and shadow on their 10:15 thumbnails, absent at 10:00): PB4 is rejected, with the reference R1; PA3_2, which stayed under the threshold, locks onto a reflection and its corr_x, corr_y make a huge jump for the time of one measurement. Everything comes back at the next measurement. The structure is not moving: this point of PA3_2 is false, to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-151", "start": "2026-10-03T10:45Z", "end": "2026-10-03T11:15Z", "family": "Obstruction", "description": "A vehicle passes on the road, headlights and bodywork translucent on the 11:00 thumbnails: PA3_2 and PA_9 are rejected for the time of one measurement, with the reference R2. The other targets do not move; everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-152", "start": "2026-10-03T11:45Z", "end": "2026-10-03T12:15Z", "family": "Obstruction", "description": "For the time of one measurement, PB1_2 and PB2_2 are rejected together: on their 12:00 thumbnails, a light silhouette passes translucent on the pavement, probably a passer-by. The other targets do not move; everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-153", "start": "2026-10-03T15:15Z", "end": "2026-10-04T15:45Z", "family": "Displacement", "description": "On 04/10, the common shift of the targets taken on 01/10 (labelled separately) fades: after the night outage, the corr_y of PA_9, PA2_2 and PB4 comes back toward its level of late September over the day, that of PB_5 on 05/10; the references PA_8 and PA1_2 come closer at the same time to the overall movement of the targets. The outages of each night prevent dating the return more precisely. Cause not established; not an established movement of the road.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-154", "start": "2026-10-03T15:30Z", "end": "2026-10-04T06:00Z", "family": "No data", "description": "No measurement for any sensor during the night, except three isolated measurements in the early morning, almost all rejected (at 05:30, all the references and all the targets, with no recalibration). The station shuts down for lack of energy: just before the outage, the battery is at its lowest and the voltage below its usual night level, and the metrics stop with the measurements; it restarts at sunrise with the solar charge, the compute board still cold (metrics). The curves are missing; nothing moves on the structure, but the station does not have enough energy to last the night: to be checked (battery, panel).", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-155", "start": "2026-10-04T04:42Z", "end": "2026-10-04T12:30Z", "family": "Thermal displacement", "description": "From sunrise until around noon, the camera moves through thermal effect, as every day: its tilt rises (the camera nods down) from the first recalibrated measurement of 04:45, already a little above its afternoon rest, to 06:15, markedly above; it goes a little below this rest in late morning (at 10:15) and regains it at 12:30. The pan drops a little at sunrise (at 06:15) then rises a little above its level (at 08:45); the rotation goes markedly above its level (at 10:15) and comes back around 12:45; the scale rises with the temperature (from 06:00 to 12:00) and comes back down in the afternoon. In the image, targets and references move together, as a block: y markedly positive at 06:15, then a little negative at 10:30 (ensemble.py --rigide; the targets move away from the block only barely). The housing temperature rises strongly from 04h to 10h and stays hot at 12h: the tilt rises with the warming of the morning and comes back while the housing is still hot; the scale follows the temperature and comes back down with it. The compensation follows: on the corrected curves, the tilt wave weighs only a little on each target; what remains is the daytime agitation, the passages and the return of the common shift (separate labels). Cause given by the expert for this station (label of 27/09): the camera moves through thermal effect, on its lens or its mount; the road does not move with temperature. Normal, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-156", "start": "2026-10-04T04:42Z", "end": "2026-10-04T15:00Z", "family": "Visibility", "description": "By day, from sunrise to mid-afternoon: the error_value of all sensors, references included, rises at sunrise and comes back down in the afternoon; the reference R3 is rejected almost all day, as every day (Target issue, separate label). The corrected curves of the targets are markedly more agitated than in the evening, without a clean step. At sunrise, the compensation loses all its references at 04:42 and 05:30, PA_9 and PB1_2 jump; PA3_2 and PB4 are rejected at 06:00 and 06:15, at 07:00, a silhouette passes in front of the reference PA1_2 (separate label); PA3_2, PA_9 and PB4 are rejected at 08:00 with the reference R2, with nothing clear on their thumbnails. Cause: daytime heat shimmer above the road and the low sun at sunrise, as every day on this station; not seen directly, the images show the scene in full daylight, the spots of the reflectors drowned in the light background. Overall compensated: the structure is not moving, do not trust the small oscillations of the day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-157", "start": "2026-10-04T06:45Z", "end": "2026-10-04T07:15Z", "family": "Obstruction", "description": "At 07:00, a dark silhouette passes translucent in front of the reference PA1_2 (07:00 thumbnail), probably a passer-by; PA1_2 and the reference PA_8 are rejected for the time of one measurement. The compensation then runs on three references (PB3_2, R1, R2; R3 is rejected as every day): the basis changes, PA2_2, PA3_2 and PA_9 jump on corr_y; everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-158", "start": "2026-10-04T08:45Z", "end": "2026-10-04T09:15Z", "family": "Obstruction", "description": "A car stops along the pavement (09:00 image) and hides the reference PB3_2 and the targets PA2_2, PA3_2, PB2_2 and PB_5, rejected for the time of one measurement; the other targets do not move. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-159", "start": "2026-10-04T10:45Z", "end": "2026-10-04T11:45Z", "family": "Obstruction", "description": "Blurry silhouettes pass on the pavement (translucent trails on the 11:00 image): the reference PB3_2 and the targets PB1_2, PB2_2 and PB_5 are rejected for the time of one measurement, PB1_2 again at 11:30. Everything comes back afterwards. Cause: a passage, probably passers-by or a vehicle; the structure is not moving, point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-160", "start": "2026-10-04T13:30Z", "end": "2026-10-04T14:00Z", "family": "Obstruction", "description": "Bright points appear near PB4 and PA3_2 (13:45 thumbnails, darker image), probably the reflections of a vehicle: PA3_2 and PB_5 are rejected, R1 and R2 set aside; PB4, which stayed under the threshold, locks onto a reflection and its corr_x, corr_y make a huge jump for the time of one measurement. Everything comes back at the next measurement. The structure is not moving: this point of PB4 is false, to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-161", "start": "2026-10-04T14:00Z", "end": "2026-10-04T14:30Z", "family": "Visibility", "description": "At 14:15, the image is very dark, with a luminous haze on the pavement near the reference PA1_2 (14:15 image): PA1_2 is rejected and the basis changes for the time of one measurement; all the targets jump together on corr_x, in the same direction, then come back at the next measurement. Cause: a failed capture (exposure, reflection), not an obstacle. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-162", "start": "2026-10-04T15:45Z", "end": "2026-10-05T06:00Z", "family": "No data", "description": "No measurement for any sensor during the night, except two isolated measurements at 04:30 and 05:00. The station shuts down for lack of energy: just before the outage, the battery is at its lowest and the voltage below its usual night level, and the metrics stop with the measurements; it restarts at sunrise with the solar charge, the compute board still cold (metrics). The curves are missing; nothing moves on the structure, but the station does not have enough energy to last the night: to be checked (battery, panel).", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-163", "start": "2026-10-05T04:30Z", "end": "2026-10-05T12:30Z", "family": "Thermal displacement", "description": "From sunrise until around noon, the camera moves through thermal effect, as every day: its tilt rises (the camera nods down) from 04:30, already a little above its afternoon rest, to 06:15, markedly above; it goes below this rest in late morning (at 11:45) and regains it at 12:30. The pan drops a little at sunrise (at 07:00) then rises a little above its level (at 08:45); the rotation goes markedly above its level (at 10:15) and comes back around 12:45; the scale rises with the temperature (around 10:00) and comes back down in the afternoon. In the image, targets and references move together, as a block: y markedly positive at 06:15, then a little negative at 11:30 (ensemble.py --rigide; the targets move away from the block only barely). The housing temperature rises strongly from 04h to 10h and stays hot at 12h: the tilt rises with the warming of the morning and comes back while the housing is still hot; the scale follows the temperature and comes back down with it. The compensation follows: on the corrected curves, the tilt wave weighs only a little on each target; what remains is the daytime agitation and the passages (separate labels). Cause given by the expert for this station (label of 27/09): the camera moves through thermal effect, on its lens or its mount; the road does not move with temperature. Normal, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-164", "start": "2026-10-05T04:30Z", "end": "2026-10-05T14:00Z", "family": "Visibility", "description": "By day, from sunrise to mid-afternoon: the error_value of all sensors, references included, rises at sunrise and comes back down in the afternoon; the reference R3 is rejected almost all day, as every day (Target issue, separate label). The corrected curves of the targets are markedly more agitated than in the evening, without a clean step. PB4 is rejected at sunrise; at 08:45 and at 11:00, vehicles pass (separate labels); at 13:30, R1 is rejected alone and PB4 and PB_5 make a small jump, with nothing clear on its thumbnail. Cause: daytime heat shimmer above the road and the low sun at sunrise, as every day on this station; not seen directly, the images show the scene in full daylight, the spots of the reflectors drowned in the light background. Overall compensated: the structure is not moving, do not trust the small oscillations of the day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-165", "start": "2026-10-05T08:00Z", "end": "2026-10-05T08:30Z", "family": "Obstruction", "description": "At 08:15, vehicle reflections dazzle the references R1 and R2 (large bright spots on their thumbnails): they are set aside, PA3_2 rejected, and the basis changes for the time of one measurement; PB4, PB_5, PB1_2 and PB2_2 jump together, then everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-166", "start": "2026-10-05T08:30Z", "end": "2026-10-05T09:00Z", "family": "Obstruction", "description": "At 08:45, as at 08:15 (separate label), the headlights of a vehicle shine near the reference R1 (08:45 thumbnail): it is rejected for the time of one measurement, the basis changes and PB4 jumps on corr_x; everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-167", "start": "2026-10-05T10:45Z", "end": "2026-10-05T11:15Z", "family": "Obstruction", "description": "A vehicle passes in front of PB4 and PA3_2, seen as translucent with its lights on their 11:00 thumbnails: PA3_2, PA_9 and PB4 are rejected for the time of one measurement. The other targets do not move; everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-168", "start": "2026-10-05T15:30Z", "end": "2026-10-06T06:30Z", "family": "No data", "description": "No measurement for any sensor during the night, then the 06:15 measurement is still missing at the restart. The station shuts down for lack of energy: just before the outage, the battery is at its lowest and the voltage below its usual night level, and the metrics stop with the measurements; it restarts at sunrise with the solar charge, the compute board still cold (metrics). The curves are missing; nothing moves on the structure, but the station does not have enough energy to last the night: to be checked (battery, panel).", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-169", "start": "2026-10-06T05:45Z", "end": "2026-10-06T12:30Z", "family": "Thermal displacement", "description": "From sunrise until around noon, the camera moves through thermal effect, as every day: its tilt (the camera nods down) is highest from the first measurement of 05:45, markedly above its afternoon rest; it comes back down quickly (a little below this rest at 07:30), makes a plateau a little above its rest from 08:00 to 09:30, when the housing temperature drops back a little, goes below its rest (at 11:00) and regains it at 12:30. The rotation goes markedly above its level in late morning (at 11:00) and comes back at 12:30; the pan is agitated around its level; the scale rises a little with the temperature and comes back down in the afternoon, at its level around 15:00. In the image, targets and references move together, as a block: y markedly positive at 05:45, then a little negative at 11:00 (ensemble.py --rigide; the targets move away from the block only barely). The housing temperature rises strongly from 04h to 07h, drops back a little at 08h and 09h, and regains its 07h level from 11h to 12h: the tilt rises with the warming of the morning and comes back while the housing is still hot; the scale follows the temperature and comes back down with it. The compensation follows: on the corrected curves, the tilt wave weighs only a little, except on PB1_2 (a little more on corr_x), whose descent from 08:30 is labelled separately; what remains is the daytime agitation and the passages (separate labels). Cause given by the expert for this station (label of 27/09): the camera moves through thermal effect, on its lens or its mount; the road does not move with temperature. Normal, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-170", "start": "2026-10-06T05:45Z", "end": "2026-10-06T15:30Z", "family": "Visibility", "description": "By day, from sunrise to mid-afternoon: the error_value of all sensors, references included, rises at sunrise and comes back down in the afternoon; the reference R3 is rejected almost all day, as every day (Target issue, separate label). The corrected curves of the targets are markedly more agitated than in the evening, without a clean step. At sunrise, PB4 is rejected with R2; at 06:45, a vehicle passes in front of PA3_2, PB4, PB_5 and the references R1 and R2 (separate label); at 08:15, when R3 comes back into the compensation, then at 09:00, when R1 leaves it, PB1_2 makes a one-measurement spike on corr_x. Cause: daytime heat shimmer above the road and the low sun at sunrise, as every day on this station; not seen directly, the images show the scene in full daylight, the spots of the reflectors drowned in the light background. Overall compensated: the structure is not moving, do not trust the small oscillations of the day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-171", "start": "2026-10-06T06:30Z", "end": "2026-10-06T07:00Z", "family": "Obstruction", "description": "At 06:45, a vehicle passes on the road above the row of targets: on the darker 06:45 image, dozens of bright points at the top of the scene (its lights and reflections, translucent), absent at 06:30. PA3_2, PB4, PB_5 and the reference R2 are rejected, R1 set aside: the compensation runs on three references (PA1_2, PA_8, PB3_2), the basis changes, PB1_2 and PB2_2 jump; everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-172", "start": "2026-10-06T07:30Z", "end": "2026-10-06T08:00Z", "family": "Obstruction", "description": "At 07:45, a passer-by walks on the pavement very close to the reference PA1_2, and a vehicle passes in the background near R1 and R2 (07:45 image): the compensation loses all its references, PA3_2 and PA_9 are rejected, and the other targets jump for the time of one measurement, PB4 the most. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-173", "start": "2026-10-06T08:30Z", "end": "2026-10-06T09:00Z", "family": "Displacement", "description": "PB1_2 alone, on the pavement kerb, goes down at once on corr_x between 08:30 and 08:45 and stays lower in the evening and the next day; its raw position has shifted a little, alone, relative to the set of targets, and its error_value does not change level. On 08/10, it goes down again, with its neighbour PB2_2 (labelled separately). Cause not established (PB1_2 had already been disturbed by something in September): a movement of the kerb is not excluded. To be checked: look at the target PB1_2, keep watching.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-174", "start": "2026-10-06T10:15Z", "end": "2026-10-06T11:15Z", "family": "Obstruction", "description": "A group of passers-by walks along the pavement and the kerb, as blurry trails in front of almost the whole scene (10:45 image): at 10:30, R1 and R2 are set aside, PA3_2 and PB4 rejected, the other targets jump; at 10:45, all the references and all the targets are rejected, the camera is not recalibrated, no curve. Everything comes back at 11:00. The structure is not moving: ignore these two points.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-175", "start": "2026-10-06T14:15Z", "end": "2026-10-06T14:45Z", "family": "Visibility", "description": "At 14:30, the image is very dark (thumbnails of R1, R2 and PB4 almost black): R1 and R2 are set aside, PB4 rejected, and the basis changes for the time of one measurement; PA3_2, PB_5 and PB1_2 jump together, then come back at the next measurement. Cause: a capture that is too dark, not an obstacle. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-176", "start": "2026-10-06T19:45Z", "end": "2026-10-07T07:00Z", "family": "No data", "description": "No measurement for any sensor during the night, except three isolated measurements at 05:00, 06:00 and 06:30, where several targets are rejected. The station shuts down for lack of energy: just before the outage, the battery is at its lowest and the voltage below its usual night level, and the metrics stop with the measurements; it restarts at sunrise with the solar charge, the compute board still cold (metrics). The curves are missing; nothing moves on the structure, but the station does not have enough energy to last the night: to be checked (battery, panel).", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-177", "start": "2026-10-07T05:00Z", "end": "2026-10-07T12:00Z", "family": "Thermal displacement", "description": "From sunrise until around noon, the camera moves through thermal effect, as every day: its tilt rises (the camera nods down) from 05:00, already a little above its early-afternoon level, to 06:30, markedly above; it goes below this level in late morning (at 11:15) and regains it at 12:00. The pan drops a little at sunrise (at 06:30) then rises a little (at 10:15); the rotation goes markedly above its 05:00 level (at 10:15) and comes back at 12:15-13:15; the scale rises with the temperature (at 07:45) and comes back down from 12:00. In the image, targets and references move together, as a block: y a little positive at 07:15, then a little less negative at 11:15 (ensemble.py --rigide; the targets move away from the block only barely). The housing temperature rises strongly from 04h to 08h and stays hot at 12h: the tilt rises with the warming of the morning and comes back while the housing is still hot; the scale follows the temperature and comes back down with it. The compensation follows: on the corrected curves, the tilt wave weighs on each target only barely more than the noise; what remains is the daytime agitation and the passages (separate labels). Cause given by the expert for this station (label of 27/09): the camera moves through thermal effect, on its lens or its mount; the road does not move with temperature. Normal, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-178", "start": "2026-10-07T05:00Z", "end": "2026-10-07T14:15Z", "family": "Visibility", "description": "By day, from sunrise to mid-afternoon: the error_value of all sensors, references included, rises at sunrise and comes back down in the afternoon; the reference R3 is rejected almost all day, as every day (Target issue, separate label). The corrected curves of the targets are markedly more agitated than in the evening, without a clean step. At sunrise, PA_9, PB4 and PB_5 are rejected at the first measurements; from 11:30 to 12:45, vehicles pass (separate labels). Cause: daytime heat shimmer above the road and the low sun at sunrise, as every day on this station; not seen directly, the images show the scene in full daylight, the spots of the reflectors drowned in the light background. Overall compensated: the structure is not moving, do not trust the small oscillations of the day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-179", "start": "2026-10-07T11:15Z", "end": "2026-10-07T11:45Z", "family": "Obstruction", "description": "At 11:30, PA3_2, PB4 and PB_5 are rejected for the time of one measurement, with the references PA1_2 and R2; the other targets barely move. A vehicle passes on the road in front of the row of targets, seen as translucent, in repeated blurry copies, on the 11:30 image. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-180", "start": "2026-10-07T11:45Z", "end": "2026-10-07T12:30Z", "family": "Obstruction", "description": "At 12:00, blurry trails pass over the references R1 and R2 (12:00 thumbnails), probably a vehicle: they are set aside, PB4 rejected, and the basis changes for the time of one measurement; PA3_2, PB_5, PB1_2 and PB2_2 jump together, then come back at the next measurement; PA3_2 is still rejected alone at 12:15, blurry translucent shapes on its thumbnail: the same passage, everything has come back at 12:30. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-181", "start": "2026-10-07T12:30Z", "end": "2026-10-07T13:00Z", "family": "Obstruction", "description": "The lights of a vehicle appear near PB4 and the reference R2 (12:45 thumbnails, absent at 12:30): PB4 and R2 are rejected for the time of one measurement. The other targets do not move; everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-182", "start": "2026-10-07T13:30Z", "end": "2026-10-08T05:30Z", "family": "No data", "description": "No measurement for any sensor during the night, the outage starting as early as the start of the afternoon, except one isolated measurement at 14:15. The station shuts down for lack of energy: just before the outage, the battery is at its lowest and the voltage below its usual night level, and the metrics stop with the measurements; it restarts at sunrise with the solar charge, the compute board still cold (metrics). The curves are missing; nothing moves on the structure, but the station does not have enough energy to last the night: to be checked (battery, panel).", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-183", "start": "2026-10-07T13:30Z", "end": "2026-10-08T13:30Z", "family": "Displacement", "description": "PB2_2, then PB1_2 again, neighbours on the pavement kerb, go down on corr_x during the outage of the night from 07 to 08/10: they are lower from the restart on the morning of 08/10, PB2_2 more markedly, PB1_2 keeps going down in the morning under the loss of the reference PA1_2 (labelled separately), and they stay lower after the return of PA1_2 (13:15) and the next day, with all the references (09/10 from 06:15), until the outage of 09/10: no return seen. This is the second step of the descent of PB1_2 from 06/10, labelled separately. Their raw positions have shifted a little, alone, relative to the set of targets; their spots stay round and sharp (08/10 16:00 thumbnail) and their error_value does not change level. Cause not established: a movement of the kerb is not excluded. To be checked: look at the targets PB1_2 and PB2_2, keep watching.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-184", "start": "2026-10-08T05:30Z", "end": "2026-10-08T13:00Z", "family": "Thermal displacement", "description": "From sunrise until around noon, the camera moves through thermal effect, as every day: its tilt rises little at sunrise (from 05:30 to 06:30-06:45, a little above its afternoon rest; markedly higher at 08:30, on a single measurement); it goes below this rest in late morning (at 10:30) and regains it at 11:45. The scale, for its part, does not rise that day. The rotation does not do its cycle: it stayed at its level until 09:00, drops markedly from 09:15 to 13:00 and holds, a little less low, in the afternoon and on 09/10, with no return; this is not the effect of the reference PA1_2, already set aside since 06:15, since the rotation stays at this low level after its return at 13:15 and with the six references at 15:00. The pan drops likewise (markedly from 07:15 to 13:15, a little less low in the afternoon). In the image, targets and references move together, as a block: y markedly positive at 08:30 (the measurement of the jump), then a little negative at 10:30; the block of targets turned like the pose between morning and afternoon: the camera did turn, and the compensation follows (ensemble.py --rigide; the targets move away from the block only barely). The housing temperature rises strongly from 04h to 10h and stays hot at 11h: the tilt rises with the warming of the morning and comes back while the housing is still hot. The compensation follows the wave: there are too few measurements with all the references to say so target by target (PA1_2 set aside almost the whole morning); on the measurements where R1, R2 and PB3_2 are retained, no more than a little; what remains is the daytime agitation, the passages, the steps of the loss of PA1_2 and the descent of PB1_2 and PB2_2 (separate labels). Cause given by the expert for this station (label of 27/09) for the daily cycle: thermal effect on the lens or the mount of the camera; the road does not move with temperature. The rotation that does not come back is not explained by the temperature (cause not established): this is not a movement of the road, the targets follow it as a block and the compensation follows it, except perhaps for PB1_2 and PB2_2, whose descent is labelled separately (link not established); the camera movement is to be watched. The daily cycle, for its part, is normal.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-185", "start": "2026-10-08T05:30Z", "end": "2026-10-08T15:00Z", "family": "Visibility", "description": "By day, from sunrise to mid-afternoon: the error_value of all sensors, references included, rises at sunrise and comes back down in the afternoon; the reference R3 is rejected almost all day, as every day (Target issue, separate label). The corrected curves of the targets are markedly more agitated than in the evening, without a clean step. At sunrise, PA3_2, PA_9, PB4 and PB_5 are rejected for nearly an hour, PA_9 until mid-morning; at 12:15 and 12:45, passages (separate labels). Cause: daytime heat shimmer above the road and the low sun at sunrise, as every day on this station; not seen directly, the images show the scene in full daylight, the spots of the reflectors drowned in the light background. Overall compensated: the structure is not moving, do not trust the small oscillations of the day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-186", "start": "2026-10-08T06:00Z", "end": "2026-10-08T13:15Z", "family": "Visibility", "description": "The reference PA1_2 is rejected almost the whole morning: on its thumbnails, the pavement in full sun is very bright and its spot is hard to make out there (08:30, 10:00), while it is sharp later (13:15). Without it, the basis changes and the rotation of the camera shifts: PB1_2, PB2_2, PA2_2 and PB_5 drop on corr_x in a step, PB1_2 and PB2_2 the most; at the brief returns of PA1_2, they rise back a little. When PA1_2 is taken again at 13:15, PA2_2 and PB_5 come back, PB1_2 and PB2_2 not quite (separate label). Cause: the daylight on the pavement around PA1_2. The steps are not a movement: curves not to be trusted over the range.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-187", "start": "2026-10-08T11:30Z", "end": "2026-10-08T12:00Z", "family": "Obstruction", "description": "At 11:45, a blurry trail passes over the references R1 and R2 (11:45 thumbnails), probably a vehicle: they are set aside, PA3_2 and PA_9 rejected, and the basis changes for the time of one measurement; PB4 and PB_5 jump on corr_x, PB1_2 and PB2_2 rise back, then everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-188", "start": "2026-10-08T12:00Z", "end": "2026-10-08T12:30Z", "family": "Obstruction", "description": "A vehicle passes in front of PA3_2, PB4 and the reference R2, seen as translucent (wheel, lights) on their 12:15 thumbnails: PA3_2, PA_9 and PB4 are rejected for the time of one measurement, with R2. The other targets do not move; everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-189", "start": "2026-10-08T12:30Z", "end": "2026-10-08T13:00Z", "family": "Obstruction", "description": "For the time of one measurement, PB1_2 and PB2_2 are rejected together, with the reference R2: on their 12:45 thumbnails, blurry silhouettes pass translucent on the pavement, probably passers-by. Everything comes back at the next measurement for these targets; the step of 13:00 comes from the reference PA1_2 (separate label). The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-190", "start": "2026-10-08T14:00Z", "end": "2026-10-08T14:30Z", "family": "Visibility", "description": "At 14:15, the image is very dark (thumbnails of R1, R2 and PB4 almost black): R1 and R2 are set aside, PA3_2 and PB4 rejected, and the basis changes for the time of one measurement; PB_5, PB1_2 and PB2_2 jump together, then come back at the next measurement. Cause: a capture that is too dark. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-191", "start": "2026-10-08T14:30Z", "end": "2026-10-08T15:00Z", "family": "Visibility", "description": "At 14:45, again a very dark image: R1 and R2 are set aside, and PB4, PA3_2, PB_5 and PB1_2 jump together for the time of one measurement, then come back. Cause: a capture that is too dark, as at 14:15. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-192", "start": "2026-10-08T16:00Z", "end": "2026-10-09T05:00Z", "family": "No data", "description": "No measurement for any sensor during the night. The station shuts down for lack of energy: just before the outage, the battery is at its lowest and the voltage below its usual night level, and the metrics stop with the measurements; it restarts at sunrise with the solar charge, the compute board still cold (metrics). The curves are missing; nothing moves on the structure, but the station does not have enough energy to last the night: to be checked (battery, panel).", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-193", "start": "2026-10-09T05:00Z", "end": "2026-10-09T12:30Z", "family": "Thermal displacement", "description": "From sunrise until around noon, the camera moves through thermal effect, as every day: until 06:00, without PA1_2, R1 and R2, the pose is distorted; afterwards, with all the references, its tilt rises (the camera nods down) from 06:15 to 07:45-08:00, later and less high than the other days, goes below the afternoon level of 08/10 in late morning (at 10:30 and 11:30) and regains it at 11:45; the measurements stop at 12:30. The pan and the rotation rise in mid-morning (the pan a little, at 10:15, the rotation markedly, at 10:30) and come back at 11:45-12:15, the rotation to the afternoon level of 08/10, not to that of before 08/10; the scale does not rise. In the image, targets and references move together, as a block: y a little positive at 06:00, then a little less negative at 10:15 (ensemble.py --rigide; the targets move away from the block only barely). The housing temperature rises strongly from 04h to 09h and stays hot at 12h: the tilt rises with the warming of the morning and comes back while the housing is still hot. The compensation follows: on the corrected curves, the tilt wave weighs only a little on each target; what remains is the daytime agitation and the passages (separate labels). Cause given by the expert for this station (label of 27/09): the camera moves through thermal effect, on its lens or its mount; the road does not move with temperature. Normal, nothing to do.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-194", "start": "2026-10-09T05:00Z", "end": "2026-10-09T12:30Z", "family": "Visibility", "description": "By day, from sunrise to mid-afternoon: the error_value of all sensors, references included, rises at sunrise and comes back down in the afternoon; the reference R3 is rejected almost all day, as every day (Target issue, separate label). The corrected curves of the targets are markedly more agitated than in the evening, without a clean step. At sunrise, the compensation loses all its references from 05:00 to 05:30: PA3_2, PA_9, PB4 and PB_5 are rejected for an hour, PA2_2, PB1_2 and PB2_2 stay measured and jump. In the morning, PA1_2 and PA_8 are set aside at times, PA_9 rejected in late morning; at 09:30, a car passes in front of PB_5 and the reference PB3_2 (separate label). The measurements stop at 12:30. Cause: daytime heat shimmer above the road and the low sun at sunrise, as every day on this station; not seen directly, the images show the scene in full daylight, the spots of the reflectors drowned in the light background. Overall compensated: the structure is not moving, do not trust the small oscillations of the day.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-195", "start": "2026-10-09T09:15Z", "end": "2026-10-09T09:45Z", "family": "Obstruction", "description": "At 09:30, a car passes in front of the bottom row, clearly seen on the thumbnails of PA3_2, PB2_2, PB4 and PB_5 (bodywork, wheel): PB_5 and the reference PB3_2 are rejected for the time of one measurement; the other targets, which stayed under the threshold, do not jump. Everything comes back at the next measurement. The structure is not moving: point to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-196", "start": "2026-10-09T11:45Z", "end": "2026-10-09T12:15Z", "family": "Obstruction", "description": "Bright vehicle reflections appear near PA3_2 and PB4 (12:00 thumbnails, absent at 11:45): having stayed under the threshold, they lock onto them and their corr_x, corr_y make a huge jump for the time of one measurement; the reference R2 is set aside. Everything comes back at the next measurement. The structure is not moving: these points are false, to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-197", "start": "2026-10-09T12:30Z", "end": "2026-10-15T08:00Z", "family": "No data", "description": "No measurement for any sensor for nearly six days. Until the early morning of 13/10, the station is mute: no metrics and no images, although the battery was not low just before (metrics of 09/10 at noon), so this is not the outage of each night. From 13/10, the metrics come back, the battery full and stable at night, and daytime images arrive every hour, but still no displacement measurement. Cause not established (station stopped, then measurement that does not restart; an intervention is possible). The curves are missing: to be checked with the operator.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-198", "start": "2026-10-15T07:45Z", "end": "2026-10-15T08:15Z", "family": "Obstruction", "description": "A single measurement, in the middle of the outage: all the references are rejected, the camera is not recalibrated, PA2_2, PA3_2, PB2_2 and PB4 are rejected; PA_9, PB1_2 and PB_5 keep a point, very far from their level. The 08:00 image shows translucent blurry trails on the pavement and the road, like the passages seen on the other days. Probable cause: a passage in front of the scene; without any reference, these points are meaningless. The structure is not moving: points to ignore.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-199", "start": "2026-10-15T08:00Z", "end": "2026-10-31T20:15Z", "family": "No data", "description": "No measurement for any sensor for more than two weeks; the metrics and the images are also missing almost all that time (two isolated values on 25/10), and come back only on 30/10 in the evening, the battery full, with daytime images on 31/10, still without a displacement measurement. Cause not established (station stopped). The curves are missing: to be checked with the operator.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-200", "start": "2026-10-31T20:00Z", "end": "2026-10-31T20:30Z", "family": "Displacement", "description": "A single measurement after three weeks without measurement: the camera has nodded down a lot since 09/10 (its tilt well above that of early October, already slowly rising from 03/10 to 09/10), all the raw positions have risen in the image. The compensation runs on five references (R2 rejected), but almost all targets have their corrected curves shifted upward on corr_y, PB1_2 and PB2_2 the most, PA_9 on corr_x, and the reprojection errors of the references are higher than usual. No image at that hour. Cause not established: the camera has tilted, and the compensation does not quite catch up with it; a single point does not say whether it is lasting. To be checked: look at the mount of the camera, keep watching the next measurements.", "review": "2026-11-04"}
{"id": "WILDSTRUBEL-201", "start": "2026-10-31T20:15Z", "end": "2026-11-01T00:00Z", "family": "No data", "description": "No measurement for any sensor until the end of the batch; the metrics stop at the same moment, although the battery was full and the voltage normal. This is not an end-of-file delay: nothing arrives on the following days either, when the other stations have their data. Cause not established (station stopped again). The curves are missing: to be checked with the operator.", "review": "2026-11-04"}