Multi-resolution Analysis Theory And Methods Of NURBS Surface Based On Anisotropic Wavelet

Curve and surface modeling module is one of the most critical part of the CAD/CAM system. For many years people are constantly exploring the convenience,flexibility and utility of new technology. Along with the development of wavelettechnology, a new breakthrough has been made in the modeling of NURBS curve andsurface with the application of this new technique, and obtained good effect. On thisbasis, the following aspects were studied.NURBS curve and surface fairing is a very important issue in CAD/CAM. In thispaper, Contourlet transform has been used for quasi-uniform B-spline surfacereconstruction in matlab7.0.This fully demonstrates its advantages in the preservation ofsurface features and the feasibility that the anisotropic wavelet can be used for NURBSsurface fairing. On this basis, thought of the existing fairing algorithm in the multi-scalefeatures in the coexistence of surface fairing insufficient, this paper points out thatanisotropic wavelet can be integrate into the multi-resolution analysis of the surfacebecause of its advantages of the high-dimensional information’s expression. Thisthought can be applied to NURBS surface fairing, in order to achieve better preservedon the surface feature.The feature points of the cross-section line contains important information aboutthe surface reconstruction. Its identification and segmentation is characteristic curveextraction, the global constraint optimization and characteristics of the correspondingskin’s premise and foundation. In view of wavelet transform in the time domain andfrequency domain with prominent local signal characteristics of ability, can reduce thenoise at the same time reserves feature of signal. Therefore, this paper presents a kind ofmethod which combines scale analysis with detail-removed wavelet reconstruction toachieve denoising of the section line and automatic identification of the feature points inthe curve fitting for multi-scale features of the coexistence of cross-sectional data.B-spline curve has been interpolated by constructing these features point to achieve theoriginal curve fitting, and fitted the curve with the maximum error control nodeinsertion method to reach the error. The experimental results show that the constructedcurve can not only remove noise, but also better retention of the original cross-sectionline characteristics...