Recognition of geometric forms using the differential depth filter

Abstract

The analysis of designs in single-piece part manufacturing during the design process results in significant benefits during the manufacturing stages. If manufacturability knowledge can be incorporated into the design, significant downstream problems may be eliminated. Manufacturing knowledge is often related to critiques on shape features that are special forms on parts that perform certain functions. The shape features required to analyze manufacturability are not always explicitly available in surface CAD models of parts, and they may need to be extracted. However, it is the variation of shape features with topology and geometry that makes the problem of feature recognition difficult. The research described in the paper focuses on shape-feature recognition that is based on the differential depth filter, which reduces the number of topological entities. A second level of operation causes the topological entities to be transformed into entities of a higher abstraction level called loops. Loops assist in reducing the number of entities in which features need to be searched for, which implies a reduction in the search space. Considerable success in feature recognition has been achieved for parts with varying topology and geometry and for parts with a relatively large number of topological entities (up to 30 000 faces and 40 000 edges).