Annotation payload

schema_version and model record which JSON shape and which ingest logic wrote the annotation — handy after you redeploy Lambda and re-ingest only some objects.

Fields the gallery uses (four filterable tags)

Field	Source	Example
camera_id	Object key	TKAH.01
utc_day	Object key	2026.170
day_phase	Filename UTC + camera timezone	dusk
visibility	Pixel brightness	daylight

day_phase values: night · dawn · day · dusk (time of capture).
visibility values: daylight · low_light · night (image brightness).

Both fields can be night independently — e.g. a dusk capture with low_light visibility.

Provenance and ops fields (stored, not emphasized in UI)

Field	Purpose
schema_version	Payload shape version ("1")
model	Ingest/classifier version (metadata-luminance-v1)
volcano_site	Site code from key
captured_utc	ISO 8601 from filename
source_bucket / source_key	GeoNet lineage
ingest_run_id	Groups one ingest invocation
ingested_at / updated_at	Write timestamps

Example JSON

Same payload shape as the CLI sample on the introduction page (get-object-annotation on 2026.170.0800.00.TKAH.01.jpg):

{
  "schema_version": "1",
  "camera_id": "TKAH.01",
  "volcano_site": "TKAH",
  "utc_day": "2026.170",
  "captured_utc": "2026-06-19T08:00:00Z",
  "day_phase": "dusk",
  "visibility": "daylight",
  "model": "metadata-luminance-v1",
  "source_bucket": "geonet-open-data",
  "source_key": "camera/volcano/images/2026/TKAH/TKAH.01/2026.170/2026.170.0800.00.TKAH.01.jpg",
  "ingest_run_id": "49f52ccc-2b61-4e1b-91c0-04440e479686",
  "ingested_at": "2026-06-19T08:57:20Z",
  "updated_at": "2026-06-19T08:57:20Z"
}

Tags are intentionally simple and verifiable — the walkthrough focus is annotate → store → query, not ML accuracy.