浅析机载LiDAR点云的滤波与分类

A Preliminary View on Filtering and Classification for Airborne LiDAR Point Clouds

滤波与分类是机载LiDAR点云处理的两个关键步骤。针对人们在传统概念上的认识误区,简要阐述了两者的不同内涵。以点云分类为例,一方面,以归一化高度特征为着力点,详细分析点云滤波对分类精度的影响,并对其影响力给予客观地评估;另一方面,以点云分类结果为着力点,采用反向分析法深入探讨了点云分类结果与滤波产生的两种误差的内在联系,从而验证了两者在点云处理过程中的辩证关系,得出了相应结论。

Filtering and classification are two critical steps in the process of LiDAR point clouds.In view of the traditional misunderstandings,their different meanings are described briefly.Taking the classification of point clouds for instance,the dialectical relation between filtering and classification is proved,and some beneficial conclusions are gained.On the one hand, based on normalized height,a detailed analysis is provided for the impact of filtering on classification,and the influences are assessed objectively;on the other hand,based on the classification result,the method of reverse analysis is developed to intensively investigate the intrinsic relationship between the classification result and the two types of errors derived from filtering of point clouds.