Policy-aware fusion

Seyond LiDAR + camera context, orchestrated by AMSYS

connecting calibration

Camera

Camera · boxes positioned by lidar, not by video

Not calibrated yet

Boxes are drawn from the lidar world frame, so the camera pose has to be solved before anything can be placed on this video. Nothing is drawn until it is, because a box in the wrong place is worse than no box.

Run calibration

Same tracks, in the room

Tracks, shared across both sensors

Nobody in the space right now. The pipeline is live, so this is an empty room rather than a dead feed.

Reprojection — Camera at — Tracks 0 Withheld as clutter —

Room view Calibrate