机载激光扫描数据粗差剔除新方法

A new method for outlier elimination of airborne LiDAR data

机载激光扫描数据的粗差剔除直接影响后期数据处理的质量,是激光扫描数据处理过程中必要的基础工作。本文在分析现有粗差剔除方法的局限性的基础上提出一种新的粗差剔除方法,利用体素实现原始激光脚点从全局到局域的转换,消除地形起伏对粗差剔除的影响,基于粗差特性分离粗差和目标激光脚点。通过实测数据实验,表明该粗差剔除方法对于任意分布的激光点云适应性强,能够对粗差进行快速定位和有效剔除,解决目前粗差剔除面临的难点问题。

Outlier elimination is a necessary step in airborne LiDAR data postprocessing. In this paper, the deficiencies of existing outlier elimination methods were analyzed,and a new method was proposed. In this method, voxel is utilized to transform the full extent raw laser points to local, which aims at avoiding the effect caused by undulant landform, and then outlier is distinguished from laser points of objects by the method proposed. According to our experiments, it appears to be adaptable in fast outlier positioning and elimination by using this method, the problems in existing methods can be solved to some extend.