A prototype detecting bottle allocations with a LiDAR camera (Realsense L515) using Python and Open3D (owlrobotics.de)
1. the floor is detected in the LiDAR point cloud (plane segmentation) and the bounding box for the floor is computed (www.open3d.org/...)
2. all point clusters for the LiDAR point cloud are computed (www.open3d.org/...)
3. the cluster with a certain distance from floor is selected as the 'bottle box' cluster - the bounding box for this cluster is computed (www.open3d.org/...)
4. from the center of the bottle box, the bounding boxes for the bottle positions are computed
5. the bottle bounding boxes are checked for points
12 сен 2024