首页 | 本学科首页   官方微博 | 高级检索  
     

增量式网格的快速对齐
引用本文:林兰芬,陈雪颂,童若锋,董金祥.增量式网格的快速对齐[J].计算机辅助设计与图形学学报,2004,16(9):1250-1255.
作者姓名:林兰芬  陈雪颂  童若锋  董金祥
作者单位:浙江大学CAD&CG国家重点实验室,杭州,310027;浙江大学人工智能研究所,杭州,310027
基金项目:国家重点基础研究发展规划项目 ( 2 0 0 2CB3 12 10 6),浙江省自然科学基金 (M 60 3 12 9)资助
摘    要:给出了一种基于增量式网格的快速对齐算法,在迭代最近点(ICP)算法的基础上,利用增量式网格多分辨率的特点,通过在不同粒度下的对齐来提高效率.还给出了一种在对齐前计算初始变化矩阵的方法,根据该方法计算出来的初始变化矩阵用于ICP算法中,可以保证网格对齐收敛于全局最佳。

关 键 词:网格对齐  增量式网格  多分辨率  初始变化矩阵  迭代最近点算法  三维扫描技术

Efficient Registration for Progressive Meshes
Lin Lanfen,Chen Xuesong,Tong Ruofeng,Dong Jinxiang.Efficient Registration for Progressive Meshes[J].Journal of Computer-Aided Design & Computer Graphics,2004,16(9):1250-1255.
Authors:Lin Lanfen  Chen Xuesong  Tong Ruofeng  Dong Jinxiang
Abstract:This paper introduces a multi-resolution-based method for the accurate and computationally efficient registration of complex triangle meshes. As an improved method of iterative closest point(ICP), our algorithm speeds up the convergence by using progressive mesh(PM) as the representation of a multi-resolution mesh with the mesh registration under different granularity. In addition, we present a new method to guarantee the registration to converge on the desired global optimism. As we know, ICP starts with two meshes and an initial guess for their relative rigid-body transformation. When the initial guess is improper, it may fall into wrong local minima. Our method can find out an initial transformation which leads to a correct registration.
Keywords:mesh registration  progressive mesh  multi-resolution  initial transformation
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号