Grain is the root of the people, and grain storage is a top priority. It is necessary to improve the efficiency of grain storage operations and reduce work intensity, so it is important to develop an automated or even intelligent grain leveling robot for the precise operation of grain silos. In this paper, we propose an area classification method based on target leveling height for the special working mode of truss-type grain leveling robot, simplify the 3D map to a 2D map, reduce the difficulty of path planning, and improve the working efficiency. Based on multi-level path planning and geneti...