Multi target tracking and dynamic scene reconstruction (back to the surveillance demo page)
Demo video: Dynamic scene reconstruction from a fixed Velodyne HDL-64E Lidar platform through multi target racking [C10]. Top left: input point cloud sequence, after applying foreground subtraction [C7], and multi target tracking. Bottom left: target tajectories from top-view. Top right: video recording of the monitored scenario (only used for reference). Bottom right: reconstructed scene with people models created in the 4D Studio [C5]. For watching the same reconstructed 4D scenario from the viewpoint of a moving simulated camera, see the following video.