A Survey of Automated Feature Recognition Techniques

Abstract

Automated feature recognition has been attempted through many methodologies for a wide range of application domains. This survey focuses on its use for geometric reasoning problems in mechanical engineering. Many of these methods are greatly limited in scope. Often they perform on a restricted class of objects with confining feature definitions. Furthermore, problems with interactions between features can render objects unrecognizable. This survey presents an overview of many of the works in this area. Included are descriptions of the approaches and an analysis of their abilities to provide a definition for and solution to the general problem of recognizing features from a solid model. It is hoped that previous research will provide the guidance for the development of a feature recognition system that is complete over a mathematically definable set of objects.