Feature recognition is a vital link in the integration of computer-aided design (CAD) and computer-aided manufacturing (CAM). The significance of automated feature recognition has increased manifold in recent years, due to the ever growing emphasis on automation of manufacturing processes. However, there are several open challenges in the field of manufacturing feature recognition, including issues of robustness, computational speed, and applicability of a given technique to different manufacturing domains.
Most of the existing research in this field employs one of the following methodologies for feature extraction: graph-based method, hint-based method and volumetric decomposition. The method proposed in this research does not exactly fit in to any of these approaches. It involves selective slicing of a given solid model at multiple levels and in multiple directions, and using contour based reasoning to identify form features. The slicing levels as well as the slicing directions are intelligently derived based on the geometry of the part. Each slice view comprises of one or more closed contours. By combining the information pertaining to the contours in successive slices, rules are formed to recognize manufacturing features.
Features that can be recognized with this method include machining and casting/molding features. Recognition of different types of intersecting features is an important advantage of this method. This technique also effectively addresses another drawback of many feature recognition algorithms, which is the inability to recognize features with varying topology and/or geometry. The proposed method is validated by successfully testing the developed rules on different 3D part models.