I love the ITS demo with lane occupancy and 3d bboxes. Is there any reason you would not use 2D detection boxes over 3d?
bump this post pls respond
hey @alex, we use 3d bounding boxes to get an accurate representation of the object dimensions, which allows us to figure out the center of the bottom plane. This is better than using just the 2D bounding box for longer vehicles or on different viewing perspectives.
1 Like