Detail feature recognition and decomposition in solid model

A methodology for abstracting features from a 3D solid model based on a new detail-level metric method is proposed. Filleting the whole boundary of an object with constant fillet radius has the effect of low-pass filtering. Taking advantage of the effect, detail-level of boundary entities can be rated. This paper investigates an approach to fillet polyhedral model and then develops a simple way to detect detailed boundary elements. Taking detailed entities as the indicators, detail features are recognized and extracted. In the detailed entities detection and decomposition cycle of the corresponding detail features, detail features are decomposed from the model one by one in terms of their locality. Detail feature decomposition directly results in geometric simplification of a 3D object. The method proposed in this paper can be applied in efficient modeling for CAE from CAD models.