Research On 3D CAD Search Technology Based On Multi-level Knowledge Representation

With economic globalization, the competition in manufacture industry turns to be more and more intensive. In order to quicken the design process, shorten the research circle, lower the cost of product development and enhance the competition of enterprise, Enterpriser and designer begin to pay more attention to how to reuse the prior model and originality. However, some factors obstacle searching model effectively from the three dimension model library, such as the different expression of date format, abstraction of key words, lacking of comprehension to the query. This paper expands researches based on the multi-level knowledge expression of 3D model. Through building ontology lexicon in 3D model domain, and giving the method to calculate the similarity of ontology based on context, we implement the multi-method 3D model retrievement. Our work mainly contains:(1) Summarizing the method to measure model character similarity:Because of largely using the similar model in the enterprise, it is very necessary to implement the shape retrieval of 3D model. In order to retrieve model effectively, this paper introduces pre-processing technologies and summarizes the methods based on model shape similarity.(2) Building ontology lexicon and calculating semantic similaritySometimes users are not very sure about the imperative models, which leads to the abstraction of key words. Therefore, in order to retrieve model effectively, we proposal the retrieval method based on ontology lexicon, which ensure the accuracy through building ontology lexicon and calculating semantic similarity based on context.(3) Developing the prototype systemMainly based on Solid Works, which is a 3-Dimension model software, and protege which is mainly used for building the ontology, our system implemented kinds of retrieval methods, including retrieve model through model character and ontology.