This project provides a ROS interface for the Microsoft Kinect One (Kinect v2) sensor, enabling robotics applications to utilize its depth, RGB, and IR data. It targets ROS users and researchers working with depth sensing and computer vision, offering a bridge to the libfreenect2 driver and tools for calibration and data visualization.

How It Works

The package acts as a bridge between the libfreenect2 library and the Robot Operating System (ROS). It leverages libfreenect2 for direct sensor data acquisition and offers optional OpenCL or CUDA acceleration for depth registration and processing. This approach aims to provide efficient, hardware-accelerated depth data for robotics tasks.

Quick Start & Requirements

• Install via git clone into your ROS catkin workspace and build with catkin_make.
• Requires ROS Hydro/Indigo, OpenCV 2.4.x, PCL 1.7.x, and libfreenect2 (>= v0.2.0).
• GPU acceleration (OpenCL/CUDA) is optional but recommended for performance.
• Launch the bridge with roslaunch kinect2_bridge kinect2_bridge.launch.
• Calibration is performed using kinect2_calibration.
• Further details and setup instructions are available in the official documentation.

Highlighted Details

• Provides a calibration tool for IR-to-RGB and depth sensor calibration.
• Includes a library for depth registration with OpenCL support.
• Offers a ROS bridge to libfreenect2.
• Features a viewer for images and point clouds.

Maintenance & Community

The project is maintained by the Institute for Artificial Intelligence at the University of Bremen. Questions and issues should be directed to the GitHub issues page.

Licensing & Compatibility

The project is available under a permissive license, suitable for commercial use and integration with closed-source ROS nodes.

Limitations & Caveats

The package is explicitly incompatible with OpenCV 3.0 due to ROS cv_bridge dependencies. Debugging information and backtraces are recommended for troubleshooting crashes. USB connection issues can manifest as transfer errors.