基于Voronoi图和Delaunay三角网的林分空间结构量化分析  

Quantitative Analysis of Forest Stand Spatial Structure Based on Voronoi Diagram & Delaunay Triangulated Network

作　　者：赵春燕[1] 李际平[1] 李建军[1] 

机构地区：[1]中南林业科技大学,长沙410004

出　　处：《林业科学》2010年第6期78-84,共7页Scientia Silvae Sinicae

基　　金：国家自然基金项目(30771724);湖南省教育厅科研项目(09C1000);中南林业科技大学青年基金项目(06004B)

摘　　要：利用Voronoi图具有最近性、邻接性特点,以单株木视为平面点构建Voronoi图和Delaunay三角网,此结构具有以下特点:对象木的最近邻木株数与Voronoi图中相对应多边形边的数目相等;对象木与最近邻木的距离和Delaunay三角网的相对应边长相等;任意2个最近相邻木的夹角即为对应Delaunay三角网2条边的夹角。在广东湛江红树林国家自然保护区选择样地进行试验,把基于Voronoi图和Delaunay三角网计算林分空间结构指数的结果与最近邻木数取4的传统算法结果进行对比分析,结果表明:两者在计算混交度、大小比数时结果有较高的相关性和一致性,但两者在计算角尺度时结果差异较大,构建Voronoi图确定最近邻木时各个方向都有选取,角尺度的取值均小于传统计算方法,但是Delaunay三角网结构本身较好反映个体在水平地面的分布格局,且基于Voronoi图和Delaunay三角网林分空间结构研究不需要对象木与邻木距离量算和角度测量。Forest stand spatial structure reflects structure associated with trees spatial position,which is generally described by mingling,neighborhood comparison and neighborhood pattern. The most important thing for calculating the three spatial structure indexes is determining the target’s nearest neighbor trees and the number of them. Voronoi diagram is one of partition patterns for spatial. This article regards individual tree as a planar point and builds Voronoi diagram and Delaunay triangulated network based on Voronoi diagram’s nearest and adjacent characteristics,which have the following features： the number of nearest neighbor trees equals the number of corresponding polygon edges in Voronoi diagrams,the distance between the target trees and the nearest neighbor trees equals the length of corresponding polygon edges in Delaunay triangulated network,and the angle between any two nearest neighbor trees equals the angle between two corresponding edges in Delaunay triangulated network. With testing in the plot in Mangrove Nation Natural Reserve at Zhanjiang,this article comparative analysises the outcomes of calculating the spatial structure indexes based on Voronoi diagram and Delaunay triangulated network and the outcomes of traditional algorithms which the number of the nearest neighbor trees is 4. The result shows： the outcomes have significantly correlation and agreement between the two when calculating mingling and neighborhood comparison, but the outcomes exist remarkably difference when calculating neighborhood pattern. Because selection is in all direction when build Voronoi diagram to determine the nearest neighbor trees,the neighborhood pattern values are less than the values of traditional algorithms,but Delaunay triangulated network structures reflect well the distribution pattern of individual on horizontal ground,and also the research of forest stand spatial structure,which based on Voronoi diagram and Delaunay triangulated network,do not need measure distance and calculate angles between the

关 键 词：林分空间结构 VORONOI图 DELAUNAY三角网 混交度 大小比数 角尺度 

分 类 号：S757.1[农业科学—森林经理学] TP18[农业科学—林学]