Skip to content

Latest commit

 

History

History
43 lines (27 loc) · 1.44 KB

README.md

File metadata and controls

43 lines (27 loc) · 1.44 KB

Lidar-Segementation

build passing velodyne_HDL_64 compliant

An implementation on "Curved-Voxel Clustering for Accurate Segmentation of 3D LiDAR Point Clouds with Real-Time Performance" from IROS 2019

Paper Link [link 1][link 2]

Update on 20210825

  1. Add a demo code
  2. Fix some problem

This file is just a function file, you may need to change a little bit to fit your own code

How to use:

 vector<PointAPR> papr;
 calculateAPR(*cloud_gr,papr);
 unordered_map<int, Voxel> hvoxel;
 build_hash_table(papr,hvoxel);
 vector<int> cluster_index = CVC(hvoxel,papr);
 vector<int> cluster_id;
 most_frequent_value(cluster_index, cluster_id);

The output is the same as https://github.com/FloatingObjectSegmentation/CppRBNN

For the demo code:

 mkdir build
 cd build
 cmake ..
 make -j

Reference

  1. Part of this code references to the https://github.com/FloatingObjectSegmentation/CppRBNN
  2. Ground Remocve: https://github.com/LimHyungTae/patchwork

Result:

Image text