Filtering of LIDAR Points by a Hierarchical Smoothing Method

Being able to collect 3D information directly from objects, LIDAR makes it possible to generate Digital Terrain Models (DTMs) quickly. To extract ground points from LIDAR point clouds, the paper proposes one filtering method which detects ground points by a hierarchical smooth method. Firstly a reference surface is generated by hierarchical smoothing method; Secondly ground points can be extracted by buffer analysis based on the reference surface. Based on ground points, DTM can be quickly produced by an interpolating algorithm. In the experiment, datasets with complex topographical undulations and different point density are taken to evaluate the validity of the proposed method. And results show that the proposed method is efficient and has potential for practices.