基于激光雷达的无人艇海上目标检测与跟踪方法研究

Research on Marine Target Detection and Tracking Method of Unmanned Surface Vehicle Based on Lidar

针对基于导航雷达和光电探测的传统感知方法难以对近距离障碍物进行稳定精确探测的问题,结合激光雷达的优势和水面场景的特点,提出了一种基于激光雷达的无人艇海上目标检测与跟踪方法。首先分析和建立了水面场景点云的平面栅格模型,提出了基于高度差的尾浪点滤除方法;其次采用改进的自适应DBSCAN算法对海上障碍物进行分割和聚类,优化了检测效果;最后结合多假设跟踪模型(MHT)和卡尔曼滤波器实现了对动态目标的多帧连续跟踪。试验结果表明,该方法能以每帧100 ms的处理速度实现对目标的实时探测和跟踪,对近距目标的聚类跟踪准确度达95%,航速、航向跟踪误差分别为8.10%和4.68%,能够实现对目标的实时、准确、稳定的检测和跟踪。

Aiming at the detection and perception problem of unmanned surface vehicle, a maritime target detection and tracking method based on lidar was proposed. Firstly, a planar raster model of the point cloud of the water surface scene was analyzed and established, and a tail wave point filtering method based on height difference was proposed. To optimize the detection effect, an improved adaptive DBSCAN algorithm was used to segment and cluster the target obstacles. Finally, the multi-hypothesis tracking(MHT) model and the Kalman filter were combined to achieve multi-frame continuous tracking of the dynamic targets. The results show that the proposed method can achieve real-time detection and tracking of targets in 100 ms per frame, the clustering accuracy for close targets reaches 95% and the speed heading tracking errors are 8.10% and 4.68%respectively, which verifies the effectiveness of the method.