In order to detect and track road targets in complex urban intersection environments,a multi-target detection and tracking algorithm based on roadside LiDAR is proposed.Firstly,the background subtraction method is used to fil-ter out the background point cloud.Then,the curved-voxel clustering algorithm is used to detect the target to obtain 3 D bounding box information with fusing 5 frame point clouds.Subsequently,a double-validation gate and life cycle man-agement strategy with adaptive threshold are put forward,which effectively improves the accuracy of object matching and reduces object missing and false detection.Finally,the fusion algorithm of Interacting Multiple Model-Unscented Kalman Filter and Joint Probability Data Association was used to track road targets.The experimental results show that the algorithm meets the real-time requirements while ensuring detection and tracking performance,and has an engi-neering application value.
关键词
激光雷达/多目标检测与跟踪/曲率体素聚类/数据关联/IMM-UKF算法
Key words
LiDAR/multi-target detection and tracking/curved-voxel clustering/data association/IMM-UKF algorithm
维普
万方数据