期刊名称：IAENG International Journal of Computer Science
印刷版ISSN：1819-656X
电子版ISSN：1819-9224
出版年度：2021
卷号：48
期号：1
语种：English
出版社：IAENG - International Association of Engineers
摘要：Obstacle detection based on 3D laser radar plays an important role on environment perception and safe operation of intelligent vehicle. The obstacle detection method based on a 3D 32-line lidar is studied in this paper. Firstly, the lidar data and the data coordinate system were calibrated. Then, road boundary suspected points were screened based on road height and smooth features. Moreover, RANSAC algorithm was used for data fitting to achieve road boundary extraction. For point clouds within the road boundary, ground and non-ground point clouds were identified by ground plane fitting method. In order to process the non-ground point clouds and realize the recognition and detection of obstacles, an improved DBSCAN algorithm was designed, which can perform clustering processing and generate the bounding boxes representing obstacles. Finally, vehicle test was carried out. Test results show that the obstacle detection method proposed in this paper can extract the road boundary accurately. Furthermore, the non-ground point cloud within road boundary and the obstacles in the non-ground point cloud were identified and distinguished.
