ingest_observation
Merge perception detections into the world model, re-identifying existing objects and updating confidence, or reconcile observed state against the expected frustum view.
Instructions
Merge structured vision perception detections into the world model (re-identifying existing objects and boosting confidence), or reconcile observed state against the expected frustum view.
Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| project | No | Optional project identifier | |
| reconcile | No | If true, also performs/returns frustum reconciliation analysis (confirmed, new, displaced, missing) | |
| detections | Yes | List of detected objects in the frame | |
| field_of_view | No | ||
| observer_pose | No | Position and orientation of the camera/agent when observing | |
| visual_state_id | No | Associated visual state ID from vision-memory-mcp |