A low-cost and accurate SLAM system that fuses Livox lidar with camera data for robust localization and mapping.
CamVox is a SLAM (Simultaneous Localization and Mapping) system that combines camera and Livox lidar data for accurate robot navigation and 3D mapping. It extends ORB-SLAM2 with lidar assistance to improve robustness, enable automatic sensor calibration, and leverage long-range depth sensing. The system is designed as a low-cost alternative to more expensive SLAM setups.
Robotics researchers, autonomous vehicle developers, and engineers working on SLAM applications who need a cost-effective, accurate system with Livox lidars. It suits those integrating visual and lidar sensors for indoor/outdoor navigation.
Developers choose CamVox for its automatic lidar-camera calibration, which reduces setup complexity, and its fusion of affordable Livox lidars with proven visual SLAM for enhanced accuracy. It offers an open-source, ROS-compatible solution optimized for real-world deployment.
[ICRA2021] A low-cost SLAM system based on camera and Livox lidar.
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
Eliminates manual sensor alignment by calibrating lidar-camera extrinsics in uncontrolled scenes, as demonstrated with static scene rosbag files triggering calibration after 10 frames of stillness.
Leverages the non-repeating scan pattern of Livox lidars for denser point clouds over integration time, improving mapping accuracy compared to traditional visual SLAM, per the animated gif and comparison figures.
Combines affordable Livox lidars with open-source ORB-SLAM2, offering a low-cost alternative to premium lidar systems while maintaining accuracy in large-scale outdoor datasets.
Fully compatible with ROS for easy deployment and testing, with provided run scripts and rosbag examples, enabling seamless hardware or simulation workflows.
Requires specific versions of Ubuntu, OpenCV, PCL, and other dependencies with manual compilation steps, making setup error-prone and time-consuming, as detailed in the lengthy prerequisites section.
Designed exclusively for Livox lidars and Hikvision cameras, limiting flexibility for teams with different sensor setups; the README links to shopping pages but offers no adaptation guidelines.
Automatic calibration only triggers in static scenes and interrupts SLAM during movement, which can be impractical for continuously operating robots in dynamic environments.