Lidar MARS Registration with optional EasyPBR

Video-Presentation | Paper

Real-time Multi-Adaptive-Resolution-Surfel 6D LiDAR Odometry using Continuous-time Trajectory Optimization
Jan Quenzel ¹, Sven Behnke ¹,
¹University of Bonn, Autonomous Intelligent Systems Group

Clone:

$ git clone https://github.com/AIS-Bonn/lidar_mars_registration.git --recursive

Install dependencies:

$ sudo apt-get install python3-catkin-pkg

Build with ROS:

$ catkin build lidar_mars_registration

Running with ROS:

$ rosrun lidar_mars_registration lidar_mars_registration_node

This runs with parameters specified in ./config/live.cfg If you want to run it with a specific parameter set (multiple defined in config subfolder), then use the following:

$ rosrun lidar_mars_registration lidar_mars_registration_node _config_file_rel:="./config/urban_loco_ca.cfg"

Make sure, that "use_ros" is set to true for the registration_adaptor.

Build with EasyPBR:

To build with better visualization, you must have first installed EasyPBR.
Afterwards this example can be build with

$ make

Running with EasyPBR:

After building one can run the executable created in the build folder with

$ ./build/temp.linux-x86_64-3.6/run_lidar_mars_registration

or

$ python3 python/registration.py

Make sure, that "use_ros" is set to false for the registration_adaptor.

Citation

@inproceedings{quenzel2021mars,
  title={{Real-time Multi-Adaptive-Resolution-Surfel 6D LiDAR Odometry using Continuous-time Trajectory Optimization}},
  author={Quenzel, Jan and Behnke, Sven},
  booktitle={Proceedings of IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
  year={2021}
}

Comparison

The modified A-LOAM and floam used within our paper comparison can be found in the following forks: floam and A-LOAM
For SuMa we first converted the bag files into KITTI binary format and used the original implementation by Behley and Stachniss: SuMa

License

We make our code available under the BSD 3-clause License.