用Delaunay三角形化实现的矩形边界表面描述算法

AN ALGORITHM FOR SURFACE DESCRIPTION CONSTRAINED BY SQUARE DOMAIN USING DELAUNAY TRIANGULATION

本文提出了一种基于Delaunay三角形化且定义在矩形边界上,具有形如Z=f(x,y)形式的表面描述算法.算法从一个简单的结构开始,在本文定义的描述误差D_K的指导下自适应地在合适的位置插入数据点以逼近实际表面,然后对旧的结构进行更新,从而获得任意精度的表面描述.对一组实际的三维物体的深度数据模拟实验表明,本算法具有程序简便,运算速度快,数据压 缩比高和存储量小的特点.

This paper presents an algorithm for surface description which is based onthe Delaunay triangulation. The surface is in the form of Z = f(x,y) and defined over a square domain. The algorithm begins from a simple structure. Under the guidance of our description difference Dk. defined in the paper, new data points are adaptively inserted in the suitable location to approximate the real surface, the old structure is refreshed, and the surface description with any required accuracy is obtained. The experiments on a set of actual range data of 3D objects show that algorithm is characterized by its simpler program structure, faster convergence speed, higher compression ratio and smaller storage space.