Uses yolo object detection and depth camera to determine position of object.
The center of the bounding box is used with a ray cast to set the tf coordinate.
The tf remains after the object is no longer detected. It uses a min_distance and eucledian distance to deduplicate. It also assigns a unique ID to the objects i.e. linear_board_0.
ros2 run object_detection yolo_to_map
Simple mapping while object is visible:
ros2 run object_detection yolo_to_pose
pip install ultralytics pip install openvino openvino-dev