关键词: 网格, 工程模型检索, 形状分布, 特征临界点, 近似测地距离, 离散曲率, 计算机辅助设计

Abstract: To effectively retrieve and reuse existing 3D model and relative design knowledge, aiming at the engineering triangular mesh model, based on mesh salient critical points, a 3D engineering retrieval algorithm was proposed, which was independent of Computer Aided Design (CAD). According to the Morse theory, the discrete mean curvature at vertices of the mesh was used as the smooth real function to calculate mesh salient critical points. Then the approximate geodesic distance between two critical points and the corresponding cosine value of the intersection angle of the two vertex-normals were employed as the United Shape Function (USF). Shape distributions were generated by computing the USF within each class of critical points, i.e. maximum, minimum, and saddle, respectively. Thus the comparison of models was mapped into the comparison of corresponding shape distribution matrixes. Experiments were conducted to evaluate the proposed algorithm by utilizing the Engineering Shape Benchmark (ESB) database developed by Purdue University. Experiential results suggested that the proposed technique effectively reflected the similarity among engineering models, and the retrieval performance was significantly improved.

Key words: grid, engineering model retrieval, shape distribution, salient critical points, approximate geodesic distance, discrete curvature, computer aided design