Extract 3D ground-truth bounding boxes from SVL
Purpose
We have 2D regions of interest in the camera images. The natural next step is to extract 3D bounding boxes emulating a perfect 3D lidar detector to also test tracking with this synthetic input.
Desired Behavior
Similar to 2D ground truth as implemented in #1099 (closed)
Definition of Done
-
A new node subscribes to 3D bounding boxes from SVL and publishes the bounding boxes and class -
Message type as defined in autoware_auto_msgs!34 (merged) is output
Edited by Frederik Beaujean