Abstract: To improve the retrieval precision with the feature modeling information in Computer Aided Design (CAD) models,a model retrieval method by taking CAD model feature topology as skeleton,features as unit and geometry of surface feature generation as detail was proposed.The secondary development technology was adopted to extract information,dependence and modeling parameters of each modeling feature,and the feature topological relation set,modeling parameter set and geometric surface set were built.The feature'based geometric surface was classified according to geometric properties,and a feature similarity calculation method was established based on surface similarity calculation.According to the feature topological relation set,the model feature topological skeleton diagram was built.The skeleton diagram was simplified by using geometric transitivity after defining the weights of features,and the similarity between the model and the retrieval one was determined by calculating the maximum common sub-graph.Experimental results showed that this method could effectively achieve the retrieval of similar CAD models with better accuracy.

Key words: computer aided design, model retrieval, similarity, feature extracting, common sub-graph

摘要： 为利用CAD模型中的特征造型信息提高检索精度,提出一种以CAD模型特征拓扑结构为骨架、以特征为单元、以特征生成曲面的几何性为细节的模型检索方法。采用CAD平台二次开发技术,从CAD模型中提取各个造型特征的约束、依赖及建模参数信息,构建特征拓扑关系集、造型参数集与几何生成面集;依据几何性质对特征几何生成面进行分类,以曲面相似性计算为基础建立特征相似性计算方法;根据特征拓扑关系集构建模型特征拓扑骨架图,在定义特征权重的基础上,利用几何传递性对骨架图进行简化,采用计算最大公共子图的方法判定检索模型与被检模型的相似度。实验结果表明,该方法能够有效实现相似CAD模型检索,具有较好的精度。

关键词: 计算机辅助设计, 模型检索, 相似度, 特征提取, 公共子图