面向点云的特征模型重构及其关键算法研究  

Research on feature model reconstruction and its key algorithm based on point cloud

作　　者：闫雒恒[1] 

机构地区：[1]华北水利水电学院信息工程学院,郑州450011

出　　处：《机械设计与制造》2012年第4期219-221,共3页Machinery Design & Manufacture

基　　金：华北水利水电学院青年科研基金资助(HSQJ2009022)

摘　　要：针对传统逆向工程技术所重构的B-rep模型信息表达层次低、无法体现设计意图、难以实现改型设计和再设计的缺陷,提出一种面点云数据直接重构特征模型的策略,并对其中点的微分几何性质估算和拉伸旋转面特征重建等基础而关键的步骤提出一套解决算法,以传统坐标转换法估算的法矢为初值,通过引入迭代来提高法矢曲率的计算精度,依据拉伸旋转面的特殊几何性质,利用随机抽样一致性算法提取拉伸方向和旋转轴,实验结果证明了算法的有效性、精确性和鲁棒性,为特征模型重构的下游计算奠定了良好的基础。Because of the low level of product information representation,B-rep model reconstructed by traditional reverse engineering is difficult to show the product designing intention,and it＇s also difficult for product redesign.To avoid the shortcoming of B-rep model,an idea of feature model reconstruction based on point cloud is presented in it.A series of algorithms,such as estimation of differential geometric properties of scattered points and extruded or rotational surface reconstruction,are also presented,which are crucial to feature modeling.An iterative process is implemented to improve estimating accuracy of normal vector and curvature,with initial normal solved by traditional coordinate transformation method being used as a starting value.According to the special geometric properties of extruded surface and rotational surface,a method is presented which can extract the extruded direction of extruded surface and the rotational axis of rotational surface with RANSAC algorithm.Experiment results verify the high efficiency,robustness and accuracy of these algorithms,which lay a good foundation for further calculation of feature model reconstruction.

关 键 词：逆向工程 点云 特征模型 重构 曲面特征 

分 类 号：TH16[机械工程—机械制造及自动化]