摘要
提出一种新的点集模型自适应增加啊采样算法.算法利用最小二乘法求出点云模型上每个点的局部光滑曲面片,并由所求得的曲面多项式计算点集曲面上每个点的曲率.通过对每个点及其邻点进行Vorono i剖分,求取每个点所控制的有效采样区域,然后根据曲率在有效区域内建立采样栅格,求取有效区域内的栅格点在曲面上的投影点即为新增采样点.该方法得到的增加采样模型可以较好地保持原点云模型曲面的几何性质,同时还可以通过选择不同的栅格得到适用于不同处理要求的点云模型.
In this paper, we propose an up-sampling algorithm for point sets. The algorithm uses the famous moving least squares method to fit local smooth patches and calculates curvature for each point in point set. Our algorithm computes local Voronoi diagram for a point with its neighbors to fred a valid upsampling region controlled by the points. Then it constructs a grid in the valid up-sampiing region according to the curvature of the point. New up-sample points can be obtained by projecting grid points in valid up-sampiing region onto the point-set surface. Our algorithm can keep geometry characters of original models and can be used in various applications to create multi-resolution models for simple point sets.
出处
《小型微型计算机系统》
CSCD
北大核心
2010年第11期2265-2271,共7页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(060873175)资助
江苏省高校自然科学基础研究项目(07KJD460108)资助
江苏省2009年度普通高校研究生科研创新项目(CX09S-009R)资助
南京师范大学优秀硕士论文培育项目资助
