Compared with the general scene

the substation scene is more complicated

and there are many similar scenes. It is not easy to be 3 D visualization. The point cloud matching based on the traditional iterative closest point（ ICP） algorithm is prone to map mismatch. In order to solve the above problems

a patrol robot system based on three-dimensional Li DAR is designed

which uses the point cloud information provided by three-dimensional Li DAR to construct a map. Meanwhile

the traditional point cloud matching algorithm is optimized

and the point cloud matching algorithm based on feature is adopted to solve the problem of mismatch

visual map deformation and location failure caused by similar scenes. The experimental results show that the method can effectively realize the three-dimensional map and prevent map deformation.