基于PROSAC算法的TLS林分树干提取研究  

Tree Trunk Extraction of the Stand Based on TLS of PROSAC Algorithm

作　　者：段勋军 刘美爽 DUAN Xun-jun;LIU Mei-shuang(Xiabeishan Forest Farm of Ledu District Haidong City in Qinghai Province,Haidong 810699,Qinghai,China;Editorial Office of Journal of Forest Engineering,Northeast Forestry University,Harbin 150040,Heilongjiang,China)

机构地区：[1]青海省海东市乐都区下北山林场,青海海东810699 [2]东北林业大学《森林工程》编辑部,黑龙江哈尔滨150040

出　　处：《西北林学院学报》2020年第1期226-231,256,共7页Journal of Northwest Forestry University

基　　金：国家重点研发计划(2017YFD0600902)

摘　　要：树干是树木的重要组成部分,对于林业调查研究具有重要意义。为了能从海量点云数据中快速精准提取林分树干,基于地面激光雷达单站点云数据,通过SOR滤波与体素化滤波对点云进行去噪和下采样处理,然后利用MLS移动最小二乘算法对数据进行平滑与孔洞修复,最后针对八叉树体元数据采用结合点法线的PROSAC算法建模提取林分树干。研究表明:PROSAC算法能适应条件复杂的林地,提取林分树干精度高、速度快、抗噪性强,可以提取多姿态树干点云,有极大的应用价值。本算法能快速精准地提取直径较大的树干点,且连续性较好,在保证算法效率与数据精度的基础上,能够获取连续性较好与姿态各异的林分树干点云。Trunk is an important part of a tree,which is of great significance to the survey and study of forestry.In order to extract forest tree trunks quickly and accurately from massive point cloud data,this study based on the ground laser radar single-site cloud data,and the point cloud was de-noised and de-sampled by SOR filtering and voxelization filtering.Then MLS was used to move the least squares algorithm to smooth the data and repair the holes,and finally,the combined point normal PROSAC algorithm was applied to model tree trunks.The algorithm could quickly and accurately extract large diameter trees,and the continuity was better.The study showed that the PROSAC algorithm could adapt to the forest land with complex conditions with high precision,fast speed,strong noise resistance.It could extract multi-attitude tree and dry cloud,demonstrating its great application value.On the basis of guaranteeing the efficiency and data accuracy of the algorithm,this algorithm could obtain the stand trunk point clouds with good continuity and different postures.

关 键 词：地面激光雷达 树干 点云 PROSAC算法 

分 类 号：S758.4[农业科学—森林经理学]