摘要
讨论了基于点云数据的曲面表示问题.采用ε-支持向量回归机和v-支持向量回归机实现点云数据的两次预处理.使用贪婪算法求解几何优化问题,从而得到点云数据的一个曲面表示.实验结果表明,所提方法具有建模光顺性好、处理速度快等优点.
Curve expression based on point cloud is discussed, ε-Support Vector Regression Machine (ε- SVRM) and v-Support Vector Regression Machine (v-SVRM) are proposed to preprocess point cloud data twice. Geometrical optimization is solved by greedy algorithm, and a curve expression based on point cloud is obtained. Experiment results show that the proposed method is of such advantages as smooth curving and fast processing speed.
出处
《信息与控制》
CSCD
北大核心
2008年第1期108-112,共5页
Information and Control
基金
国家自然科学基金资助项目(10576027)
关键词
支持向量机
点云
子采样
曲面表示
support vector machine (SVM)
point cloud
sub-sample
curve expression