With the rapid increase in the number of three-dimensional (3D) models each year, to quickly and easily find the part desired has become a big challenge of enterprises. Meanwhile, many methods and algorithms have been proposed for part retrieval. However, most of the existing methods are designed for mechanical parts, and can not be well worked for sheet metal part retrieval. An approach to feature-based retrieval of sheet metal parts is presented. Firstly, the features frequently used in sheet metal part design are chosen as the "key words" in retrieval. Based on those features, a relative position model is built to express the different relationships of the features in 3D space. Secondly, a description method of the model is studied. With the description method the relative position of features in sheet metal parts can be expressed by four location description matrices. Thirdly, based on the relative position model and location description matrices, the equivalent definition of relationships of two feature groups is given which is the basis to calculate the similarity of two sheet metal parts. Next, the formula of retrieval algorithm for sheet metal parts is given. Finally, a prototype system is developed to test and verify the effectiveness of the retrieval method suggested. Experiments verify that the new method is able to meet the requirements of searching sheet metal parts and possesses potentials in practical application.
展开▼
