基于车载激光点云的数学形态学滤波方法改进

针对车载激光点云数据空洞插值误差大及利用多尺度数学形态学滤波方法耗时高的问题,提出一种改进的数学形态学滤波方法,该方法按照点云空间分布形态格网化,以控制开运算方向。腐蚀运算过程中,根据格网腐蚀前后高程的变化情况,更新格网高程,以达到空洞精确填充及各格网均含地面点的目的。利用5组点云数据进行了实验,结果表明:该方法能有效提取地面点,运算效率明显优于多尺度数学形态学滤波方法,具有一定的实用性。

Morphological Filtering Method Improvement Based on Vehicle-Borne Laser Point Cloud Data

In order to overcome interpolation error of vehicle-borne laser point cloud data and time-consuming problem of multi-scale morphological filtering method.An improved method of filtering is proposed.This method need gridding according to the spatial distribution of point cloud to control the direction of opening operation.Updating grid elevation to fill the hole according to elevation changes during erosion operation.Five groups of point cloud data is tested.The result shows that the filtering method can remove various types of ground objects,and more effective than the multi-scale morphological filtering method in calculation.