Abstract: In order to overcome the large number of misclassification errors caused by the general automatic filtering algorithm for complex terrain, the paper proposes a line classification algorithm for LiDAR data based on slice data. The algorithm uses a priori knowledge of the terrain of human judgment and multi-level automatic filtering method and high threshold to obtain the initial filtering result. Then the angle characteristics of 3D space are used to optimize the result. Finally we can achieve a good filtering effect on the breakline terrain. In this paper we use the VC++ programming to realize the algorithm of line classification, and prove the applicability of the algorithm through experimental analysis and comparison, which can adapt to the rapid production of high precision DEM.

Key words: airborne LiDAR, point cloud, DEM, filter, line classification