This package is a collection of GICP-based fast point cloud registration algorithms. It constains a multi-threaded GICP as well as multi-thread and GPU implementations of our voxelized GICP (VGICP) algorithm. All the implemented algorithms have the PCL registration interface so that they can be used as an inplace replacement for GICP in PCL.

• FastGICP: multi-threaded GICP algorithm (~40FPS)
• FastGICPSingleThread: GICP algorithm optimized for single-threading (~15FPS)
• FastVGICP: multi-threaded and voxelized GICP algorithm (~70FPS)
• FastVGICPCuda: CUDA-optimized voxelized GICP algorithm (~120FPS)

on melodic & noetic

• PCL
• Eigen
• Sophus
• nvbio
• OpenMP (optional)
• CUDA (optional)

We have tested this package with Ubuntu 18.04, ROS melodic, and CUDA 10.2.

To enable CUDA-based features, uncomment find_package(CUDA) in CMakeLists.txt.

cd ~/catkin_ws/src
git clone https://github.com/SMRT-AIST/fast_gicp --recursive
cd .. && catkin_make -DCMAKE_BUILD_TYPE=Release

git clone https://github.com/SMRT-AIST/fast_gicp --recursive
mkdir fast_gicp/build && fast_gicp/build
cmake .. -DCMAKE_BUILD_TYPE=Release
make -j8

CPU:Core i9-9900K GPU:GeForce RTX2080Ti

roscd fast_gicp/data
rosrun fast_gicp gicp_align 251370668.pcd 251371071.pcd

target:17249[pts] source:17518[pts]
--- pcl_gicp ---
single:116.732[msec] 100times:10867.1[msec] fitness_score:0.204306
--- pcl_ndt ---
single:52.8007[msec] 100times:5220.49[msec] fitness_score:0.226416
--- fgicp_st ---
single:110.343[msec] 100times:10651.2[msec] 100times_reuse:6962.1[msec] fitness_score:0.0922969
--- fgicp_mt ---
single:24.3643[msec] 100times:2716.7[msec] 100times_reuse:1799.1[msec] fitness_score:0.0922969
--- vgicp_st ---
single:115.041[msec] 100times:8759.43[msec] 100times_reuse:4784.57[msec] fitness_score:0.0912174
--- vgicp_mt ---
single:19.705[msec] 100times:1963.74[msec] 100times_reuse:1044.29[msec] fitness_score:0.0912174
--- vgicp_cuda (parallel_kdtree) ---
single:16.1846[msec] 100times:1611.89[msec] 100times_reuse:779.65[msec] fitness_score:0.0709287
--- vgicp_cuda (gpu_bruteforce) ---
single:49.7294[msec] 100times:3145.78[msec] 100times_reuse:1541.36[msec] fitness_score:0.0710122

See src/align.cpp for the detailed usage.

# Perform frame-by-frame registration
rosrun fast_gicp gicp_kitti /your/kitti/path/sequences/00/velodyne

• ndt_omp
• fast_gicp

• Kenji Koide, Masashi Yokozuka, Shuji Oishi, and Atsuhiko Banno, Voxelized GICP for fast and accurate 3D point cloud registration [link]

Kenji Koide, [email protected]

Robot Innovation Research Center, National Institute of Advanced Industrial Science and Technology, Japan [URL]