velodyne_track_person

ROS node for tracking a person in point clouds from 3D LIDAR. The task is: given the data from a Velodyne sensor on a mobile robot, there is need to detect a person moving in the room.

Test data

Bag files for testing algorithm are awailable here.

Usage:

Run ROS master:

roscore

Play rosbag:

rosbag play --clock velodyne_1_person.bag

Run rviz:

rosrun rviz rviz

Running velodyne_track_person node:

rosrun velodyne_track_person velodyne_track_person <options>

Options are:

• _h=true: Show help
• _res:= Octree resolution for the Octree change detector (side length of octree voxels)
• _noise:= Noise filter for the Octree change detector
• _mean_k:=<mean_k> Number of points to use for mean distance estimation in StatisticalOutlierRemoval noise filtering
• _std_dev_mul:=<std_dev_mul> Standard deviation multiplier threshold for StatisticalOutlierRemoval noise filtering
• _rad:= Correspondence grouping size
• _min_neighbors:=<min_neighbors> Minimum number of neighbors for RadiusOutlierRemoval noise filtering