The Application And Research Of Subdivision Wavelet For The Processing Of Three-dimensional Mesh Data In Reverse Engineering

With the advent of the age of intelligentize industry, the reverse engineering technology has already become an effective and necessary approach that the modern enterprises can complete the technological innovation, accelerate the product update and meet the demand of market quickly. Whether in military enterprise or daily product processing, the complement between reverse technology and 3-D printing technology is benefit for the innovation and development of industrial products. In this technology, how to realize the preliminary de-noising, smoothing, restructuring and design again of three-dimensional physical model through the data obtained and processed, that is the key point of design and research of reverse engineering, and it is benefit for more precise formation to put into the market.In this process, the application of using three-dimensional laser scanning to obtain the 3-D meshes of archetype is more common. In this paper, the experiment combining with the three-dimensional laser scanner EXAscan has accomplished. The three-dimensional graph and the large number of irregular triangle grids data of experimental model which needs to process has obtained by scanning original model and it has established the foundation for subsequent analysis.De-noising processing is the essential step before leading the meshes into reverse software, and the wavelets technology and subdivision method have outstanding contribution on multi-resolution analysis of Image and simplification and compression of meshes, combining the two technology and leading them into the industry can speed up the research process of key step. But, the Calculation process of combining the two methods simply is too complicated, a reasonable method should be selected and to be improved for the application. This paper utilizes the subdivision wavelets that constructed by the combination of the subdivision principles and wavelet to process the model. In the treating, choosing the Loop subdivision wavelet which adapts to the details de-noising, and it carries on the improvement of the auto-adapted apex importance which is the apex deleting of Loop subdivision wavelets. Through the data analysis on the platform MATLAB software and writing improved algorithm procedure, detail de-noising experiment of massive triangle grid data has been realized.After processing in MATLAB, the de-noised meshes model obtained by this method has used to compare with the model data de-noised with original algorithm on effect. And it has also compared with the model de-noised through the Geomagic Studio software, the contrast value can be obtained from aspects separately, such as deviation, time and simplicity of methods. Ultimately concluded, it proves that the de-nosing method using Loop subdivision wavelet with improvement is more convenient, accurate and quick than the method without improvement and the reverse software.