多约束的平面点集形状重构方法

针对平面点集空间分布的复杂性,本文提出了一种基于Delaunay三角网的平面点集形状重构方法。首先采用一种简单且实用的数据结构以表达Delaunay三角网中嵌入的几何信息和拓扑信息,然后由外向内迭代过滤Delaunay三角网得到一个大概边界,最后进一步考虑边界的凹凸信息和空洞现象,获取最终的精细边界。试验结果表明与其他典型的Delaunay三角网重构方法相比,本文提出的算法能更好地适用于平面点集空间分布的复杂性,通过所构建的数学模型实现了凸凹多边形内外边界提取。

An Efficient Approach to Shape Reconstruction from Planar Point Set Based on Multi-constraints

An efficient algorithm to boundary representation from a planar point set in order to adapt the complexity of spatial distribution was presented in this paper.At first,an appropriate and practical data structure was designed to express geometric information and topological information,which provides an easy access to links embedded in DT serving as a basis for the filtering procedures;then the algorithm generates rough boundary based on an iterative removal of Delaunay triangulation.Furthermore,a mathematic formulation for cavities and holes was given and a statistical method to detect them was designed.Finally,a series of experiments including both simulated and real data sets to validate the effectiveness and practicability of our algorithm was conducted.