Skip to content

Latest commit

 

History

History
79 lines (59 loc) · 3.21 KB

missing_metadata_for_usgs.md

File metadata and controls

79 lines (59 loc) · 3.21 KB

Missing flight metadata — request list for USGS

We're missing *_captures.csv (and the corresponding *_flights.csv, *_cameras.csv, and .aflight) for the flights below. Everything else we thought was missing turned out to be a naming issue on our side and needs nothing from you.

Our metadata lives in metadata_aflight_csvs/ (230 flights) and metadata/aflights/. For each flight below we have the imagery but no metadata record of any kind.

1. Fully missing flights (4 flights, 133,969 images)

Flight key Images Notes
20241219_120500 20,964
20241219_131500 37,942
20241219_150200 4,544
20260202_141900 70,519

These are partial-day gaps, which is what makes us think it's an incomplete transfer rather than data that was never collected:

  • 2024-12-19 — we received 160100, 164400, 165600, 172400. We are missing 120500, 131500, 150200 (the earlier flights that day).
  • 2026-02-02 — we received 094800, 122400. We are missing 141900 (the last flight that day).

Nothing for these four exists on our side: no captures/flights/cameras CSV and no .aflight file.

2. Incomplete captures coverage (1 flight, ~5,000 images)

2023-04-25 (SE_JPG_2023_April25_NC): we have 5,812 images, but only 817 of them (14%) appear in any captures CSV.

We do have 20230425_094400, 095500, 095800, 104500, 135600, 140400. The unmatched images are captured between roughly 10:00 and 11:59 — the 104500 captures file exists but contains only 1 of our ~5,159 images from that hour, so it looks truncated or belongs to a different camera/flight than we assumed.

Could you check whether the captures export for the 2023-04-25 late-morning flight(s) is complete?

3. Minor residuals (not urgent, ~180 images total)

Small numbers of images don't appear in any captures CSV. Probably not worth chasing, listing for completeness:

  • JPG_March29b (2024-03-29): 74 of 4,712 unmatched
  • 2023_Sept5_jpgs: 54 of 3,141 unmatched
  • 2023_Sept7_jpgs: 27 of 787 unmatched
  • JPG_2024_Jan27: 24 of 1,018 unmatched

Not a USGS problem (for our records)

The other 14 "missing metadata" folders were a false alarm. Their images are in the captures CSVs we already have — the folder names just don't encode a flight datetime, so our lookup failed. Joining on the image Basename resolves them 100%:

Folder Resolves to
JPG_2021_March24 20210324_135000
2023_June20b_JPG 20230620_130800
2023_June21_800ft_JPG 20230621_123800
JPG_2023_Dec14 20231214_114800
JPG_2024_Jan5_6b 20240105_104600
JPG_2024_Jan31 20240131_095100, 20240131_115500
JPG_2024_Jan31b 20240131_132100
JPG_2024_Feb29a 20240214_150400
JPG_2024_Feb29b 20240214_112500
SE_JPG_2023_April24_NC 20230424_* (6 flights)

Note the folder names are unreliable: JPG_2024_Feb29a actually contains imagery captured on 2024-02-14. The image filename (C1_L10_F3295_T20240214_161903_292.jpg) is the source of truth, joined to the captures Basename column.

Also screened_images/JPG_20241219__150200 has a double underscore (typo) — same flight as 20241219_150200 above.