Plane Seg

Plane Seg PointNet2 is a library for fitting planes to noisy depth camera data. It is developed based on (https://github.com/ori-drs/plane_seg). The main idea is to use PointNet2 for plane points classification, and then cluster the plane points to planes.

Using It

python env: This package requires python 3.7.

conda env create -f cupy.yml

Run-Time Application: A sample application can be run and visualized in Rviz as follows:

conda activate cupy
source /opt/ros/noetic/setup.bash
catkin_make -DPYTHON_EXECUTABLE=/home/minghan/mambaforge/envs/cupy/bin/python
source devel/setup.bash 
roslaunch plane_seg_ros view_plane_seg.launch

Test program: reads example point clouds (PCDs), processes them and executes the fitting algorithm: In the first terminal:

conda activate cupy
source devel/setup.bash 
roslaunch plane_seg_ros cw_test2.launch

In the second terminal:

conda activate cupy
source devel/setup.bash 
roslaunch plane_seg_ros cw_test3.launch

In the third terminal:

conda activate cupy
source devel/setup.bash 
roslaunch plane_seg_ros cw_test1.launch