Feature Recognition

Feature Recognition

The term "feature" does not imply the same meaning in different engineering disciplines. This has resulted in several ambiguous definitions for feature. A feature, in computer-aided design (CAD) software, can be called a region of a part with some interesting geometric or topological patterns. This meaning can refer to all sorts of information, such as for example, shape, functional or manufacturing information. Although many types of features have been investigated, the most common type of feature is the form feature, which contains both shape information and parametric information. Examples of form features common in many shape models are round holes, slots, bosses, and pockets.

Features can also be used to represent manufacturing information of the part. Different manufacturing domains require different feature representations. Some of the properties that need to be encoded by features are assembly method, manufacturing process and tolerances. A manufacturing feature can be defined as a form feature, but not necessarily vice versa. Among manufacturing features, the ones received extensive attention are the machining features. A machining feature can be regarded as the volume swept by a cutting tool. In this sense, it is always a negative (subtracted) volume, in contrast with form features that are sometimes positive (added) volumes.

Feature data in a CAD model can be represented either as a collection of surfaces or volumetrically. Surface features are naturally used for example to describe manufacturing tolerances or locating surfaces in fixture design. volumetric features on the other hand, are used in process planning since manufacturing information (particularly in machining) is better portrayed volumetrically.

The first published work on features was for the original boundary representation modelling system, BUILD, and was performed by Lyc Kyprianou. Soon other work followed based on different solid representations. Overviews on the work on features can be found in Shah et al.; Subrahmanyam and Wozny; Salomons et al.