Feature recognition from solid models using conceptual shape graphs
Sheng-Hsiung Frank Chuang · 1991
In this research, a method is presented for both computer recognition of shape features and subsequent feature-based decomposition of a solid from a boundary representation (B-rep). Labelled graphs called conceptual shape graphs are used to represent shape concepts such as features which are used to extract the shape concept from a B-rep solid model. First, the B-rep of a solid is preprocessed and transformed into a shape graph for feature recognition and decomposition. A shape graph is the face-edge graph extracted from a B-rep with labelled shape elements. Shape elements are edges and faces classified according to their types and degrees of convexity. The shape graph is the input for the feature searching and decomposition algorithms. A simple feature is defined by a feature graph, which is conceptualized from a regional shape of a solid with regards to a specific application. A simple feature is recognized by finding a subgraph from a shape graph in which the subgraph is isomorphic to the feature graph. Because of the high complexity in subgraph matching, a node classification algorithm is introduced to reduce the search space. A compound feature, defined by a graph grammar, is a group of regional shapes. A compound feature is recognized by parsing the shape graph of a solid according to the corresponding graph grammar. The recognition process decomposes the surface of a solid into a collection of features. The feature relationships for the solid are represented in a graph where the features are the nodes and their relationships are the arcs. The definitions of features requiring only application knowledge, and the recognition process, pure mathematical searching technique, are clearly separated. The advantages are that features are both easy to define and conform to a mathematical formalism.