凸壳内缩法进行多密度离散点群边界检测  

Boundary detection of multi-density point cluster using convex hull retracted method

作　　者：李雯静[1] 李少宁[1] 邱佳[1] 周侗[2] 

机构地区：[1]武汉科技大学资源与环境工程学院,武汉430081 [2]南京师范大学地理科学学院,南京210046

出　　处：《测绘科学》2014年第9期126-129,共4页Science of Surveying and Mapping

基　　金：国家自然科学基金(41271449;41171350);高校博士点基金(20103207110012)

摘　　要：针对现有离散点群边界检测方法对多密度点群检测结果效果不够理想这一问题,本文提出利用凸壳内缩法检测离散点的边界:先建立离散点群凸壳结构,再定义内缩精度后通过点群凸壳结构特征实现最适内缩精度的识别。该方法对内缩扫描区域面积会随边界点密度不同而变化,因此更加适用于多密度离散点群边界的检测。通过与现有边界检测方法的对比,本文提出的方法在处理多密度离散点群的边界检测问题时具有高效性和普适性。Aiming at the unsatisfactory effect of boundary detection algorithm on multi-density point cluster, this paper presented a method using convex hull retracted. At first, convex hull of the multi-den- sity point cluster was established, and then a definition of retracting accuracy, which was used to deter- mine the precise extent of the point cluster＇s boundary, was proposed according to characteristic of the convex hull. For the scanning area of retracting differs by the densities of point cluster, the algorithm in this paper would be applied in boundary detection of multi-density point cluster better. Comparative experi- ment of alpha-shape algorithm and angle algorithm showed the feasibility of the method.

关 键 词：边界检测 凸壳内缩 多密度离散点群 内缩精度 

分 类 号：P208[天文地球—地图制图学与地理信息工程]