dodeakim/nano_gicp

Nano-GICP as a module from the official github repo: [IEEE RA-L & ICRA'22] A lightweight and computationally-efficient frontend LiDAR odometry solution with consistent and accurate localization.

Nano-GICP: Fast-GICP + Nano-FLANN

• This branch is for being used as a module in other packages
• Nano-GICP is from here, official repo of DLO or DLIO

Dependencies

• PCL >= 1.8
• C++ >= 14
• OpenMP >= 4.5
• CMake >= 3.10.0
• Eigen >= 3.3.7

Use case

• Refer - here
• Or, refer to the example usage as follows:
  1. catkin build this repository in your workspace,
  2. In the CMakeLists.txt of your wanted package, import nano_gicp as a component of catkin

  find_package(catkin REQUIRED COMPONENTS
    ...
    nano_gicp #Include here
    ...
  )
  include_directories(
    ...
    ${catkin_INCLUDE_DIRS} #Header files are included in catkin_INCLUDE_DIRS
    ...
  )
  add_library(some_library src/some_src.cpp)
  target_link_libraries(some_library ${catkin_LIBRARIES}) #Libraries are included in catkin_LIBRARIES

  3. Use in the source file of your wanted package as:

  #include <nano_gicp/point_type_nano_gicp.hpp> //change PointType in this headerfile, currently pcl::PointXYZI
  #include <nano_gicp/nano_gicp.hpp>
  
  nano_gicp::NanoGICP<PointType, PointType> m_nano_gicp;
  
  ////// nano_gicp init
  m_nano_gicp.setMaxCorrespondenceDistance(max_corres_dist_);
  m_nano_gicp.setNumThreads(thread_number_);
  m_nano_gicp.setCorrespondenceRandomness(correspondences_number_);
  m_nano_gicp.setMaximumIterations(max_iter_);
  m_nano_gicp.setTransformationEpsilon(transformation_epsilon_);
  m_nano_gicp.setEuclideanFitnessEpsilon(euclidean_fitness_epsilon_);
  m_nano_gicp.setRANSACIterations(ransac_max_iter_);
  m_nano_gicp.setRANSACOutlierRejectionThreshold(ransac_outlier_rejection_threshold_);
  
  ////// use
  pcl::PointCloud<pcl::PointXYZI>::Ptr src_(new pcl::PointCloud<pcl::PointXYZI>);
  pcl::PointCloud<pcl::PointXYZI>::Ptr dst_(new pcl::PointCloud<pcl::PointXYZI>);
  pcl::PointCloud<PointType> dummy_;
  /* watch out! */
  *src_ = src_data; //put your data here
  *dst_ = dst_data; //put your data here
  /* watch out! */
  m_nano_gicp.setInputSource(src_);
  m_nano_gicp.calculateSourceCovariances();
  m_nano_gicp.setInputTarget(dst_);
  m_nano_gicp.calculateTargetCovariances();
  m_nano_gicp.align(dummy_);
  
  double score_ = m_nano_gicp.getFitnessScore();
  // if matchness score is lower than threshold, (lower is better)
  if(m_nano_gicp.hasConverged() && score_ < icp_score_threshold)
  {
    Eigen::Matrix4f pose_betweenf_ = m_nano_gicp.getFinalTransformation(); //float
    Eigen::Matrix4d pose_betweend_ = m_nano_gicp.getFinalTransformation().cast<double>(); //double
  }

License and acknowledgements

This work is licensed under the terms of the MIT license.

• FastGICP - Kenji Koide, Masashi Yokozuka, Shuji Oishi, and Atsuhiko Banno, “Voxelized GICP for Fast and Accurate 3D Point Cloud Registration,” in IEEE International Conference on Robotics and Automation (ICRA), IEEE, 2021, pp. 11 054–11 059.
• NanoFLANN - Jose Luis Blanco and Pranjal Kumar Rai, “NanoFLANN: a C++ Header-Only Fork of FLANN, A Library for Nearest Neighbor (NN) with KD-Trees,” https://github.com/jlblancoc/nanoflann, 2014.
• DLO - Kenny Chen, Brett T. Lopez, Ali-akbar Agha-mohammadi, and Ankur Mehta, “Direct LiDAR Odometry: Fast Localization With Dense Point Clouds,” in IEEE Robotics and Automation Letters, 2022.