Abstract: LiDAR point cloud data has the problem of large data volume, not easy to be recognized and difficult to deal with. In order to solve the above problems,point cloud data needs classification and processing. Aiming at the problem that the accuracy of point cloud classification is not high and the complexity of the process is complicated,a new method of point cloud classification based on height difference is proposed,which is combined with TerraSolid software.Using the method to separate the point cloud data,the first step is to deal with the original LiDAR point cloud data by TerraSolid software to remove the noise and extract the surface point cloud,and then use the second derivative of regular building and irregular vegetation height difference to extract which may be the point of the building or vegetation,and use the Gaussian deviation estimation model to provide the threshold for the classification of the building and vegetation,and finally use the breakpoint statistical model to supplement the building and vegetation point cloud.In order to prove the feasibility and effectiveness of this method,the point cloud classification of LiDAR point cloud data in Autzen_Stadium region is used.The results show that the method has the advantages of good feasibility,good sorting effect and processing automation.

Key words: LiDAR point cloud data, point cloud data classification, elevation difference, building, vegetation