特征CAD模型的全局对称的快速识别方法

An Approach to Rapidly Detect the Global Symmetries in Feature-based CAD Models

针对现有面向CAD模型的对称识别方法效率低下的问题,提出一种充分利用特征信息,通过特征分类、特征集排序、合并和验证对称,快速识别特征CAD模型的全局反射对称和旋转对称的方法.通过分析特征信息中隐含的对称信息,可以快速获取特征的对称;根据对称强弱关系的原理,利用特征集排序使得中间结果快速向目标对称靠拢;合并和验证对称能避免检测不必要的局部对称,以快速获取全局对称,并能较好地解决根据特征信息无法识别对称以及避免从几何上识别对称的问题.实验结果证明,采用文中方法能正确且快速地识别特征CAD模型的全局反射对称和旋转对称.

The existing methods have a low performance of detecting the symmetries in CAD models. Therefore, an approach is proposed to rapidly detect the global reflectional and rotational symmetries in feature-based CAD models by making full use of feature information. In detail, the symmetries of CAD models can be derived by classifying features, sorting feature sets, merging and verifying symmetries of the feature sets. The symmetries of the features can be easily obtained by using feature information. According to the principles of the weak/strong relationships in symmetries, the intermediate detected symmetries can move quickly to the target symmetries of the entire model. Merging and verifying symmetries avoid unnecessary detecting of all the local symmetries to derive the global symmetries. And also, the problems that the symmetries can not be detected by using feature information and how to avoid geometrically detecting symmetries can be efficiently solved. Experimental results show that the global symmetries can be correctly and rapidly detected.