1.change the "root dir name" in "main.py" to the absolute path of your velodyne data, and run "main.py"
python main.py
2.project including ground segemantation, downsampling, clustering, and a optimized version of kdtree.
a) split tree along dimension with max variance instead of so called "round axis" or random axis
b) sorting part of data in each dimension
