The Study On CAD Mesh Model Segmentation By Clustering

CAD mesh models have been widely employed in current CAD/CAM systems. It is quite useful to recognize the features of the CAD mesh models in the applications, such as automated process planning, sheet metal tool design, high-quality finite element analysis (FEA) mesh generation, reverse engineering, and redesign/reuse of feature shapes for product design, etc. The first step of feature recognition is to segment the CAD mesh model into meaningful parts. Although there are lots of mesh segmentation methods in literature, the majority of them are not suitable to CAD mesh models, as those models usually contain triangles with very different sizes and shapes. The aim of this dissertation is to study and implement an effective mesh segmentation method, which is dedicated to the CAD mesh models and can automatically distinguish different surfaces in the model well.By virtue of the characteristics of the CAD mesh models, a mesh segmentation method based on clustering is proposed in this dissertation. Specifically, by clustering the shape information with the agglomerative hierarchical clustering method, the triangles in the given CAD mesh model is first classified into sparse triangles and dense triangles. Furthermore, the parts with sparse triangles are segmented into planar regions, cylindrical regions, and circular conical regions by the Gauss map of the triangular faces and Hough transformation. Meanwhile, the dense regions in the model are also segmented by the mean shift clustering method performed on the mean curvature field defined on the mesh faces. The method presented in this dissertation can automatically segment a CAD mesh model into meaningful parts, while the user can also improve the result by controlling several key parameters.Based on the study above, a CAD mesh model segmentation system is implemented in this dissertation. Lots of empirical results demonstrate the effectiveness and efficiency of the CAD mesh segmentation method proposed in this dissertation.