Automatic Recognition of Boolean Segmentation Loops for B-Rep Models
Zhang Yingzhong, Luo Xiaofang, and Zhao Yong
(School of Mechanical Engineering, Dalian University of Technology, Dalian 116024)
Boolean segmentation loops are the traces of design features modeled on B-rep models, and their automatic recognition is the foundation for the subsequent reconstruction of design features. Based on the shape evolution analysis of the part models in the feature-based modelling process, an approach to automatic recognition for Boolean segmentation loops is presented. This approach first constructs a feature vertex adjacent graph of the part model to be reconstructed according to the convex-and-concave properties of geometric edges and the adjacency properties of geometric vertexes, and implements searching for the vertex node of segmentation loops. Then, the search procedure is divided into four steps, namely the initial node selection, the node growth, searching evaluation, and feedback operations. In the node growth step, a decision method in terms of the minimum distance to the feature surface and a method of forming pseudo edges by virtually linking adjacency vertexes are employed, which can ensure the segmenting loop closed. The experimental results show that this method can automatically recognize Boolean segmentation loops from the complex intersecting edges formed by intersection of design features.