Research on method of path planning of unmanned surface vehicle based on LiDAR

To address the inefficiency and inaccuracy in water surface target detection and path planning for unmanned surface vehicles (USVs), this paper proposes a water surface target detection method based on LiDAR and a safe-region RRT* path planning algorithm based on KD-Tree. First, environmental point cloud data collected by LiDAR is processed using a bilateral filtering algorithm to filter out scattered and disordered noise points, thereby reducing the amount of data for subsequent processing. Next, the improved RANSAC algorithm is applied to achieve water surface segmentation, distinguishing between water and non-water regions. Subsequently, a multi-threshold Euclidean clustering algorithm is used to complete water surface target detection and identify water surface obstacles. Based on this, a path planning algorithm is designed to generate a safe and feasible navigation path for the USV. Simulation and experimental results demonstrate that the proposed method exhibits high reliability and effectiveness in water surface target detection and path planning, providing essential technical support for the autonomous navigation of USVs in complex water environments.