关键词: 三维CAD模型, 局部检索, 图谱理论, 索引, 过滤

Abstract: To tackle the inefficiency problem of the current partial retrieval methods for 3D CAD models,a new partial retrieval approach based on graph indexing and filtering mechanism was proposed.By extracting B-Rep information such as geometry and topology attributes of the models,3D CAD models were represented by Attribute Adjacency Labelled Graphs (AALGs).By employing the spectral graph theory,AALGs and their vertices were encoded respectively.Based on these spectral graph codes,the lemmas were proposed and proved which gave the essential condition on partial matching of CAD models,and a two-level filtering strategy combined with a graph code indexing was presented according to these lemmas.A filter-and-verification framework was constructed to speed up the partial retrieval for 3D models.Experimental results showed that the proposed method could speed up the efficiency of partial retrieval significantly,which satisfied the requirements of engineering practice.

Key words: 3D CAD model, partial retrieval, spectral graph theory, indexing, filtering