A Robust Ground Point Cloud Segmentation Algorithm Based on Region Growing in a Fan-shaped Grid Map
- Resource Type
- Conference
- Authors
- Liu, Zhenbo; Yi, Zhenhui; Cheng, Changwei
- Source
- 2022 IEEE International Conference on Robotics and Biomimetics (ROBIO) Robotics and Biomimetics (ROBIO), 2022 IEEE International Conference on. :1359-1364 Dec, 2022
- Subject
- Computing and Processing
Robotics and Control Systems
Point cloud compression
Biomimetics
Semantics
Clustering algorithms
Autonomous aerial vehicles
Robustness
Robots
Ground point cloud segmentation
over-segmentation
region growing
fan-shaped grid map
- Language
In order to reduce the over-segmentation ratio in ground point cloud segmentation under Unmanned Aerial Vehicle(UAV) platform, a new algorithm based on fan-shaped grid map and region growing is proposed. At first, the concept of the lowest ground point of a grid in fan-shaped grid map is introduced which is obtained by Euclidean clustering and the lowest ground point with lowest height in the map is used to initialize the region growing algorithm. Then all ground points are obtained by region growing based on gradient and height constraints associated with the lowest ground point. Finally, a semantic grid mapping system is developed based on the above ground segmentation algorithm to verify the effectiveness and robustness of our algorithm in practical applications. The experimental results show that the algorithm can correctly judge the ground area under UAV platform and effectively overcome the problem of over-segmentation, and the mapping results show that the proposed ground segmentation algorithm is robust and practical under UAV platform. We open source our implementations at https://github.com/liuzhenboo/mav_find_road.