图匹配特征识别的实现

Graph-Based Matching for Recognition of Machined Features

提出了用图匹配理论实现特征识别的基本原理 ,研究了判定边的凸凹性、属性邻接图、设计特征等理论与方法 。

The CAD system models the shape of a workpiece i n computational form. Machined feature recognition can convert those geometric a nd topologic information into high-level information required for manufacturing .. This paper proposed a matching method to identify sub-graph(feature) from th e attribute adjacency graph(AAG), AAG involves arcs, attributes and nodes which g ive the constraints of topological relationships, geometric relationships and a number of concerned entities. It can be decomposed into several sub-graphs on m atching rules. The sub-graphs accurately matching with the template-feature w hich is precisely defined in the designed feature library can be recognized. We developed the concept of attribute adjacency graph, the concave or convex attrib ute of edge and feature definitions. It can be used to recognize the feature fro m medium complex model.