docs(perception): record Rerun projection acceptance
This commit is contained in:
@@ -45,6 +45,22 @@ of the existing result.
|
||||
## Interpretation
|
||||
|
||||
This accepts the external recorded compute boundary, not detector quality or a
|
||||
vehicle decision loop. The result has not yet been projected into Rerun. SSH/SCP
|
||||
is retained only as laboratory bootstrap evidence and is not the product data
|
||||
plane.
|
||||
vehicle decision loop. SSH/SCP is retained only as laboratory bootstrap evidence
|
||||
and is not the product data plane.
|
||||
|
||||
## Rerun projection acceptance
|
||||
|
||||
Mission Core subsequently revalidated the exact job/result generation,
|
||||
reconstructed and decoded the 56-frame camera epoch, and serialized one complete
|
||||
RRF2 overlay (approximately 62.55 MB) for the active TEST007 recording ID. A
|
||||
second render of the same recording/result identity reused the private cache.
|
||||
The HTTP endpoint returned the exact declared length, RRF2 magic and
|
||||
`application/vnd.rerun.rrd` without exposing storage paths.
|
||||
|
||||
The Control Station accepted the overlay on an isolated Rerun channel. The
|
||||
operator-facing **Распознавание** switch appeared only after acceptance and
|
||||
selected the native `/perception/camera` Rerun view. Browser QA at the shared
|
||||
session timeline displayed the recorded left-camera image with a green
|
||||
`person · 34%` box; **Облако точек** returned to the spatial view. The tested
|
||||
result is camera-only: no point-cloud model, image/point fusion, tracking,
|
||||
free-space estimate or live inference was executed.
|
||||
|
||||
Reference in New Issue
Block a user