Research And Application On Entity Feature Information Extraction

Feature is a bridge of CAD/CAPP/CAM and a good connector of geometry topology information on CAD and concrete engineering information. It heightens application of CAD, CAPP and CAM and product models make new sense from geometry topology information to special engineering information.According to the basic theory on feature information, this thesis deep researches feature and feature recognize and develops a system of feature abstraction using the method of cross-sectional profile automation feature recognition. This system can provide primary data for later program.This thesis discusses the development of feature recognition technology and analyzes deficiencies on it according to product model. By defining and classifying feature and analyzing the main feature recognition methods, it emphatically researches the theory on graph-based feature recognition and brings forwards arithmetic on feature recognition, the arithmetic can figure out lacks of intelligent on extraction feature information.Universal work-pieces are main 2.5D models in machine. In allusion to their characteristics, this thesis deep discusses concepts involved in the course of graph-based feature recognition and analyzes recognition theory on the isolated feature. This thesis intensively researches feature recognition from the perspective of theory. Feature recognition method is respectively analyzed in detail about the data structure of discrete point and generic data structure, accordingly it is elicited that a mature degree of constituted matching regulation determines a degree of identified feature. Aim at 2.5D parts principal features, a suited element and a basic arithmetic about limbic concave and protruding are constructed. Mainly suited elements are quantified and an end face loop matrix is built. On the basis of it, a method of feature information recognition and extraction is put forward, by which feature information can obtain in some surface of an entity part, and it must match with preconcerted regulations, in order to obtain its geometry and topology information.A feature recognition system on entity parts is developed according to before-mentioned researches. This system can pick up single feature on a work-piece, obtain its absolute space position's dimension and process information from process data-base. The information will be provided with following process.