树木枝干点云数据的等值线提取  

Contour extraction of point cloud data for tree branches

作　　者：高士增[1] 张怀清[1] 刘闽[1] 何清平 罗立平 

机构地区：[1]中国林业科学研究院资源信息研究所,北京100091 [2]湖南省攸县黄丰桥林场广黄分场,湖南攸县412306

出　　处：《浙江农林大学学报》2013年第5期648-654,共7页Journal of Zhejiang A&F University

基　　金：国家高技术研究发展计划("863"计划)项目(2012AA102002);国家林业公益性行业科研专项(201104028);国家自然科学基金资助项目(31170590)

摘　　要：通过三维激光扫描技术得到的树木枝干点云数据,不仅数据量大,而且特征复杂,不适合采用传统的方法提取等值线。对此,首先把点云数据中树木的枝和干分为不同的部分,然后建立点云的分层模型,并分析点云在树高方向上的数据量分布。在高精度采样下,将分层点云作为等值线的采样数据,对每层数据中不同部分的树木枝干点云分别采用凸包算法进行连接,建立树木枝干的等值线模型。结果表明:在没有先验等值线知识和建立点云对象模型的条件下,利用迭代的凸包算法可以有效地对树木枝干离散点云数据进行连接,得到的等值线符合一般等值线的特点。最后通过实例验证了方法的适用性。3D laser scanning technology was used with point cloud data of tree branches to measure the inherent complex characteristics and contour model of the branches was established.Branches were divided into different parts,and a hierarchical model of the tree was established.Also,distribution of the point cloud for tree height was analyzed.With high-precision sampling,the stratified point cloud was used for contour sampling data,and all points of different parts of the branches for each floor of data were connected using a convex hull algorithm.Then a contour model of the branches was established followed by validation of the model.Branches diameter of different height were extracted from the model,the error was less than 5％ compared with the actual measurement value.Results showed that in the absence of prior knowledge of the contours and with the establishment of the point cloud objective model,an iterative convex hull algorithm effectively connected discrete point cloud data on the tree branches,and the model was validated.Thus,with the new contours being consistent with the original contours,feasibility of this approach was verified.

关 键 词：森林测计学 三维激光扫描 树木枝干 等值线 凸包算法 

分 类 号：S758.5[农业科学—森林经理学]