AUROVA-LAB / aurova_detections

aurova_detections

yolo_ros_ouster

This package implements deep learning methods to track pedestrians using images reconstructed from LiDAR point clouds. It implements two trackers, one based on YOLO and other in DaSiamRPN.

m2track_ros

This package implements the point cloud tracker $M^2$-Track in ROS.

tracker_filter

This package fuses the predictions of the trackers implemented in yolo_ros_ouster and m2track_ros using an Extended Kalman Filter. It also publish a point cloud without the target, so the local planner can follow the target without considering it an obstacle.