Comparisons of adaptive TIN modelling filtering method and threshold segmentation filtering method of LiDAR point cloud

8th International Symposium of the Digital Earth

地球科学;计算机科学

Chen, Lin^1,2 ; Fan, Xiangtao^1 ; Du, Xiaoping^1

Key Laboratory of Digital Earth, Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences, No.9 Dengzhuang South Road, Haidian District, Beijing 100094, China^1

University of China Academy of Sciences, No.19A Yuquanlu, Beijing 100049, China^2

关键词: Different terrains;    Filtering method;    Height parameters;    Irregular networks;    Lidar point clouds;    Parameter selection;    Threshold segmentation;    Topological relations;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/18/1/012028/pdf
DOI  :  10.1088/1755-1315/18/1/012028

学科分类：计算机科学（综合）

来源: IOP

【 摘 要 】

Point cloud filtering is the basic and key step in LiDAR data processing. Adaptive Triangle Irregular Network Modelling (ATINM) algorithm and Threshold Segmentation on Elevation Statistics (TSES) algorithm are among the mature algorithms. However, few researches concentrate on the parameter selections of ATINM and the iteration condition of TSES, which can greatly affect the filtering results. First the paper presents these two key problems under two different terrain environments. For a flat area, small height parameter and angle parameter perform well and for areas with complex feature changes, large height parameter and angle parameter perform well. One-time segmentation is enough for flat areas, and repeated segmentations are essential for complex areas. Then the paper makes comparisons and analyses of the results by these two methods. ATINM has a larger I error in both two data sets as it sometimes removes excessive points. TSES has a larger II error in both two data sets as it ignores topological relations between points. ATINM performs well even with a large region and a dramatic topology while TSES is more suitable for small region with flat topology. Different parameters and iterations can cause relative large filtering differences.