Research On Recognition Of Features In Semantic Feature Medeling

Feature modeling as one of the new technologies of CAD/CAM integrated system which is the key of product design. Built a feature-based, unified and complete product information model is necessary, it could implement the maintenance and editing of model. Semantic feature modeling is a declarative feature modeling approach. It is not only able to provide detailed specification and well-defined semantics, but also can maintain the semantics of the whole feature effectively in the modeling process. The shape information and functional information of feature modeling combine together to constitute the semantics of features. But the existing feature modeling techniques are mostly aimed at the regular shapes, increasingly practical applications related to freeform feature, and it may bring about many problems when manipulating such features with the existing methods, so this is important problem to be solved.Feature recognition is an approach of extracting machining features from entity design, which constitutes an ideal interface between CAD and CAPP. Over years of research and development, scientists proposed many kinds of efficient automatic feature recognition methods.Aimed at the shortcomings of the traditional feature recognition, this paper analyzed the modeling process relating to feature recognition process, put forward the parameterized definition of feature, researched template matching recognition method of freeform feature. This paper improved the limits of this method, proposed template-based evolution of freeform feature recognition method. Firstly, gave the parameterized definition of freeform feature and analyzed the completeness of the definiton. Then according to the definition, put forward the parameterized definition of feature recognition method and the general rule of the effectiveness of method. Secondly, on the basis of the existing template matching recognition method, re-gave the step of the method and validity analysis with the parameterized thinking. Aimed at the shortcoming of template matching method, improved it and proposed template-based evolution of freeform feature recognition method. This method simulated the evolution of biological thought, first identified the corresponding feature in the feature library, which is the best matching with the target feature. Then by the two evolutionary processes—the parameter values evolution and the parameter mapping evolution, got the most suitable corresponding feature on the target base surface. By the feature-based parameterized deformation, separated the base surface of target feature. Finally, from the first best matching feature, started a new evolutionary feature recognition process, and achieved the final result of feature recognition.