Files
aimsight-labeling/auto/analyses/NORDEND_2026-10-24T1630_2026-11-07T1400/report.md
T
maximeandClaude Opus 5.5 ddb17c2189 auto: labels requalified under the 5-noise rules (only what stands out from the noise): 32 analyses refined, the 6 paused periods verified, the long labels requalified with the fixes the period analysts found; 1153 -> 961 labels
- 32 analyses: each label kept, rewritten or removed by its analyst (section Requalification of the report; mm magnitudes where the targets have them)
- SNCF01 x3, WILDSTRUBEL 2026-09-01 -> 2026-10-06 x3: verified from their existing labels under the new rules
- long labels (no analysis): auto/analyses/long-labels/<STATION>/; NORDEND-169 and -195 run to 07/11, so NORDEND-283 and -284 (their continuation) are removed
- reading-rules.md, stations.md: the citations of removed or rewritten labels point to their successor, or to [expert]

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-17 23:58:00 +01:00

7.6 KiB
Raw Blame History

NORDEND: the labels of 24 October to 7 November, kept to what stands out from the noise

You asked to requalify the labels of 24/10 16:30 → 07/11 14:00 (UTC) under the new rules: only what stands out from the noise of the corrected curves is labelled.

In short

28 labels stay, 8 are removed: single missed points that moved nothing else, a jolt too small to stand out, and one thermal day the correction absorbed. The curves can't be trusted during the rain, the fog, the low sun, the nights when the cobweb covers reference CTR03, or around midday on most sunny days. PR09, PR10, PR11 and PR17 stay offset by a few mm after the rain. No real movement of the quarry is seen.

Requalification

Kept: 26 unchanged. Rewritten: 2. Removed: 8. The deviation of a target is counted in its own noise; the threshold is 5.

  • NORDEND-281 (28/10 11:45, PR04 missed once): removed. Only a missing point, on a target that doesn't enter the correction. The small dip of other curves at that time belongs to the thermal cycle of the day.
  • NORDEND-286 (02/11, thermal cycle): removed. The correction absorbs it: no target goes beyond 4.9 noises from its night level.
  • NORDEND-289 (03/11 13:30, camera jolt): removed. At most 3.6 noises (PR01, vertical).
  • NORDEND-232 (05/11 morning, PR09 rejected at times): removed. Only missing points, and PR09's kept points stay within 3.9 noises. The spike at 03:30 keeps its own label (CTR03).
  • NORDEND-235, NORDEND-238, NORDEND-241, NORDEND-242 (05/11 15:00 PR09, 06/11 09:30 PR10, 07/11 11:30 PR04, 07/11 14:00 PR07): removed. Each is a single missing point; no target moves beyond 5 noises.
  • NORDEND-218 (26/10) and NORDEND-220 (27/10), thermal cycles: rewritten. Their text said the correction absorbs the movement, but the corrected curves keep it: PR18 horizontal at −7 and −10 noises, PR10 vertical at −9 noises on 27/10.

What happened

Rain, then fog, then low sun (24–25/10, 27–28/10, 30/10). On the night of 24 to 25 October, rain drops on the window hide most targets and the three references in turn. The remaining curves jump by up to about 40 mm (PR18 vertical). On the morning of 28 October, fog makes the correction fail for one measurement at 07:15: about 30 to 40 mm on many targets. On 30 October around 09:30, a low sun leaves the quarry in shadow: about 25 mm. Confirmed on the images.

Drop marks left on the window (from 25/10 to the end). After the rain, marks stay on the window near PR09, PR10, PR11 and the reference CTR02. These three targets have stayed offset ever since, by about 3 to 4 mm (PR09 horizontal, PR10 vertical), 10 to 13 times their noise. PR17 keeps the offset left by an older mark in front of it: about 2 mm horizontal against early October, about 6 mm vertical against September. Marks seen on the night images; that they cause the offsets is likely.

The daily thermal cycle of the camera (every day except 2 November). The camera turns a little each morning as its housing warms, and comes back in the evening. The correction absorbs most of it, not all: around midday the targets on the right edge drift, mostly PR18 (116 m away), by about 2 to 4 mm horizontally. That is 7 to 16 times its noise. From 1 to 3 November the housing stayed warm at night and the camera did not come back until the evening of the 3rd. Confirmed thermal: the camera follows the housing temperature. Why the nights stayed warm is unknown: there is no weather record for this station.

The camera position from 31 October to 5 November Top panels: the pan and the tilt stay on a plateau from the 1st to the 3rd instead of making a daily wave.

The cobweb (2–7 November). Threads of a web in front of the lens, lit by the infrared, cross the group of targets every night. On the night of 2 to 3 November, a knot sits on PR07 and PR08: PR07's kept points slide by about 2 mm, PR08's by about 1 mm. On 5, 6 and 7 November, a thread covers reference CTR03. Each time it is lost, the correction runs on two references and almost all curves jump together for that measurement: about 7 to 8 mm vertically on PR17 and PR18. Confirmed on the images.

The cobweb at 22:00 on 2 November The bright lines are the web threads; the knot at the bottom of the group of dots sits on PR07 and PR08.

Other moments. At 12:00 on 5 November, a one-measurement jolt of the camera leaves a dip of about 2 mm on several curves (PR01, PR10). The cause is not established. CTR01, the local reference read as a target, is rejected two times out of three, every day. Three short gaps (26, 27 and 29 October) have no measurement at all. Other stations had the same gaps, so the cause was in the transmission, not the station.

What to do

Clean the window: the drop marks, the mark in front of PR17 and the cobweb are all on it. Check CTR01 (its spot is barely visible by day). Nothing to do for the camera movements.

Not established

  • Why the housing stayed warm at night from 1 to 3 November: daytime images or a weather record would tell whether warm air or cloud was there.
  • Whether PR17's present level is still the mark or a return: a check of the prism and a clean window would settle it.

Labels

Start (UTC) End (UTC) Family Targets In a few words
24/10 16:30 25/10 09:00 Visibility almost all, CTR02–04 night of rain
25/10 05:45 07/11 14:00 Visibility PR09, PR10, PR11 drop marks on the window, lasting offsets
25/10 07:30 26/10 07:15 Thermal displacement PR01, PR12, PR18 weak camera cycle
26/10 07:15 27/10 06:30 Thermal displacement PR18, PR17 weak camera cycle, no return at night (rewritten)
26/10 14:15 26/10 15:00 No data all common gap
27/10 06:30 27/10 16:30 Thermal displacement PR18, PR10, PR13 camera cycle (rewritten)
27/10 08:45 27/10 09:15 No data all common gap
27/10 16:30 28/10 09:15 Visibility most, references degrading night, fog
28/10 09:15 29/10 06:45 Thermal displacement PR18, PR15–17 camera cycle
29/10 06:45 30/10 07:45 Thermal displacement PR18, PR16, PR10 camera cycle
29/10 14:45 29/10 15:15 No data all common gap
30/10 07:45 31/10 07:45 Thermal displacement PR18, PR16, PR10 camera cycle
30/10 08:45 30/10 10:45 Visibility most, CTR02, CTR03 low sun
31/10 07:45 01/11 06:00 Thermal displacement PR18, PR16, PR10 camera cycle
01/11 00:00 07/11 14:00 Target issue CTR01 rejected two times out of three (now carried by NORDEND-195, extended to 07/11)
01/11 00:00 07/11 14:00 Obstruction PR17 window mark, lasting offset (now carried by NORDEND-169, extended to 07/11)
01/11 06:00 02/11 06:30 Thermal displacement PR18, PR13 camera turned, no return at night
02/11 19:15 03/11 06:00 Obstruction PR07, PR08 cobweb thread
03/11 07:00 03/11 23:00 Thermal displacement PR10, PR01–PR12 camera comes back as the housing cools
03/11 23:00 05/11 06:30 Thermal displacement PR17, PR18 camera cycle of 04/11
05/11 00:45 05/11 01:30 Obstruction CTR03, all cobweb on a reference
05/11 03:15 05/11 03:45 Obstruction CTR03, all cobweb on a reference
05/11 06:30 06/11 07:00 Thermal displacement PR18, PR17 camera cycle
05/11 11:45 05/11 12:15 Displacement PR01, PR10 camera jolt
05/11 23:00 06/11 10:00 Obstruction CTR03, all cobweb on a reference
06/11 07:00 07/11 06:30 Thermal displacement PR18, PR16 camera cycle
06/11 21:15 07/11 11:00 Obstruction CTR03, all cobweb on a reference
07/11 06:30 07/11 14:00 Thermal displacement PR18, PR06 camera cycle