城区LiDAR点云数据的树木提取  

Extracting Trees from LiDAR Data in Urban Region

作　　者：张齐勇[1] 岑敏仪[1] 周国清 杨晓云[1,3] 

机构地区：[1]西南交通大学土木工程学院测量工程系,四川成都610031 [2]Department of Engineering and Technology,Old Dominion University [3]广西工学院土木建筑工程系,广西柳州545006

出　　处：《测绘学报》2009年第4期330-335,共6页Acta Geodaetica et Cartographica Sinica

摘　　要：机载激光扫描(LiDAR)可以快速获取地球数字表面模型。提出一种适合复杂城市环境的机载激光扫描数据提取树木的算法:首先对LiDAR数据滤波生成DTM,提取地物点;然后对地物点进行区域增长运算,使用面积阈值滤出大的区域;再计算出LiDAR数据点的梯度值,根据梯度阈值分离出树木点;最后结合梯度阈值分割和区域增长分割的结果实现树木点的最终提取。实验结果表明,使用新算法在城区环境中能从LiDAR数据中较好地提取出树木,城区树木提取率达到85.4%,提取正确率为86.1%。Light Detection and Ranging(LiDAR)can acquire digital surface model quickly.As an important portion of city surface information,the tree is important to the development of 3D digital city.This paper presents a tree extracting algorithm from LiDAR data in complicated urban environment.Firstly,a DTM is generated from LiDAR data,non-ground points are extracted subsequently.Then,the algorithm tempts to remove conspicuous building area,using region-growing algorithm based on area threshold value.As following,the...

关 键 词：LIDAR 树木提取 区域增长 梯度阈值分割 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]