Compare Map Filter

Autoware package that compare the LiDAR PointCloud and PointCloud Map and then extract (or eliminate) coincident points.

Requirements

  • PointCloud Map with extremely few unnecessary PointCloud (people, cars, etc.).

How to launch

  • From a sourced terminal:
    • roslaunch points_preprocessor compare_map_filter.launch
  • From Runtime Manager:

Sensing Tab -> Points Preprocessor -> compare_map_filter You can change the config, as well as other parameters, by clicking [app]

Parameters

Launch file available parameters:

Parameter Type Description
input_point_topic String PointCloud source topic. Default /points_raw.
input_map_topic String PointCloud Map topic. Default /points_map.
output_match_topic String Comparing LiDAR PointCloud and PointCloud Map, topic of the PointCloud which matched. Default /match_points.
output_unmatch_topic String Comparing LiDAR PointCloud and PointCloud Map, topic of the PointCloud which unmatched. Default /unmatch_points.
distance_threshold Double Threshold for comparing LiDAR PointCloud and PointCloud Map. Euclidean distance (mether). Default 0.2.
min_clipping_height Double Remove the points where the height is lower than the threshold. (Based on sensor coordinates). Default -2.0.
max_clipping_height Double Remove the points where the height is higher than the threshold. (Based on sensor coordinates). Default 0.5.

Subscribed topics

Topic Type Objective
/points_raw sensor_msgs/PointCloud2 PointCloud source topic.
/points_map sensor_msgs/PointCloud2 PointCloud Map topic.
/config/compare_map_filter autoware_msgs/ConfigCompareMapFilter Configuration adjustment for threshold.
/tf TF sensor frame <-> map frame

Published topics

Topic Type Objective
/match_points sensor_msgs/PointCloud2 Comparing LiDAR PointCloud and PointCloud Map, topic of the PointCloud which matched.
/unmatch_points sensor_msgs/PointCloud2 Comparing LiDAR PointCloud and PointCloud Map, topic of the PointCloud which unmatched.

Video

Compare Map Filter Autoware