Obstacle_processor

ROS package for obstacle segmentation in a point cloud scene.

Installation

• Before you install the package, you have to configure your RGB-D sensor and calibrate it. You can also run this package offline i.e. streaming point cloud ROS topic from .bag file.
• Clone the repository inside src/ directory of your catkin workspace

mkdir obstacle_processor
cd obstacle_processor/
git clone name_of_repository

• Run CMake to compile source code

catkin_make

• Source your workspace

source catkin_ws/devel/setup.bash

• Setup your robot platform on a ground and remove the all objects in front of it for calibration purposes and run calibration node

roslaunch obstacle_processor calibration.launch

• Now you can run obstacle_processor detection algorithm by either of 5 launch commands (two last commands launch obstacle_processor_node along with kinect2_bridge package from iai_kinect2 package, but can be replaced for whatever bridge package compatible with your RGB-D sensor that produces point cloud ROS topic)

roslaunch obstacle_processor obstacle_processor.launch

or

roslaunch obstacle_processor obstacle_processor_rviz.launch

or

roslaunch obstacle_processor obstacle_processor_rviz_debug.launch

or

roslaunch obstacle_processor obstacle_processor_launch_all.launch

or

roslaunch obstacle_processor obstacle_processor_launch_all_rviz.launch

References

The project was done as a part of research during bachelor thesis

License

MIT

Citation

If you use the repo in personal project or research, please cite it as follows:

RYBIN, A. Detekce překážek za použití kamerového 3D skeneru. Brno: Vysoké učení technické v Brně, Fakulta strojního inženýrství, 2018.

Copyright © 2017 Andrei Rybin