Obstacle Detection and Modeling Based on Lidar Scan Lines

When a mobile robot is driving on the road, the detected target motion status and matching association are the key to the intelligent robot’s dynamic obstacle tracking. This paper uses multi-line mechanical rotating lidar as the main sensor to conduct research on obstacle detection and modeling. When establishing the obstacle data model, this paper proposes a clustering labeling algorithm based on scanning lines based on the working scanning principle of lidar. It analyzes the horizontal and vertical distance resolution of lidar and is based on the improved DBSCAN clustering algorithm. Based on the dynamic value selection principle of the mid-distance threshold, experimental results show that this algorithm can effectively reduce the computational cost of the clustering algorithm and improve the clustering effect of the algorithm.