| During the past decade, automatic recognition of design and machining features has been an active area of research in computer-aided design and manufacturing and many new algorithms have been proposed. Most algorithms for feature recognition in the literature require three dimensional geometric models as input. However, these algorithms are complicated and not very efficient for the extraction of intersecting features. Besides, companies worldwide have accumulated through years of engineering practice, vast amounts of two-dimensional drawings containing bulk of the information on past designs. Researchers believe that methods to obtain three dimensional feature based models from these legacy drawings will continue to be of significant economic interest to these companies.; The International Organization for Standardization (ISO) first introduced the standard for product definition for process planning using machining features, STEP AP224, in 1999 and later revised it in December 2000. However, recent research in the area of machining feature recognition has not considered the feature details as given in the updated standard.; This thesis proposes and implements a methodology to identify interacting machining features directly from two dimensional drawings and obtain AP224-defined machining features. The methodology is based on the ideas in rule-based and delta volume decomposition approaches to recognize features in 3D. The algorithms proposed in the thesis take as input two-dimensional drawings in a neutral file format, identify through/blind features and produce the corresponding 2.5D data structures that can be used to build 3D parametric, feature-based models. |
