ros2_pcl_segmentation

ROS2 Package for point cloud segmentation using PCL library. This repository contains multiple segmentation algorithms for point clouds.

• pcl_car_segmentation: Package for segmenting cars from a point cloud.
• pcl_ground_segmentation: Package for segmenting ground from a point cloud.
• pcl_human_segmentation: Package for segmenting humans from a point cloud.

Get Started

Installation

cd ~/ros2_ws/src
git clone https://github.com/CDonosoK/ros2_pcl_segmentation.git
git submodule update --init --recursive
rosdep install -i --from-paths ros2_pcl_segmentation --rosdistro rolling -y
colcon build --packages-select ros2_pcl_segmentation

Prerequisites

To test the package, you can use the kitti dataset. You can download the dataset from the following link: KITTI Dataset or follow the instructions below:

cd mkdir ~/ros2_ws/data
cd ~/ros2_ws/data
git clone https://github.com/Deepak3994/Kitti-Dataset.git 
bash raw_data_downloader.sh

Then to convert the dataset to a rosbag file, you will need the following package:

cd ~/ros2_ws/src
git clone https://github.com/umtclskn/ros2_kitti_publishers.git
colcon build --packages-select ros2_kitti_publishers

From Docker

You can also run the package using docker. You can use the following command to build the docker image that is inside the docker folder:

docker build -t ros2_pcl_segmentation .

Then you can run the docker image as follows:

sh run_docker.sh

Usage

For each ROS2 Package I have created a launch file that will run the segmentation algorithm. You can run the launch file as follows:

• pcl_car_segmentation:

  ros2 launch pcl_car_segmentation bring_kitti.launch.py
  ros2 launch pcl_car_segmentation bring_rviz.launch.py

• pcl_ground_segmentation:

  ros2 launch pcl_ground_segmentation bring_kitti.launch.py
  ros2 launch pcl_ground_segmentation bring_rviz.launch.py
  

• pcl_human_segmentation:

  ros2 launch pcl_human_segmentation bring_kitti.launch.py
  ros2 launch pcl_human_segmentation bring_rviz.launch.py