A strategy for B-spline curve data reduction based on non-uniform B-spline wavelet decomposition is presented. In existing methods of knot removal, ranking the removal knots depends on a procedure of assigning a weigh...A strategy for B-spline curve data reduction based on non-uniform B-spline wavelet decomposition is presented. In existing methods of knot removal, ranking the removal knots depends on a procedure of assigning a weight to each knot to indicate its significance. This is reasonable but not straightforward. Propose is a more straightforward and accurate method to calculate the weight. The wavelet coefficient is taken as a weight for the corresponding knot. The approximating curve and the error can be obtained directly from the wavelet decomposition. By using the hierarchical structure of the wavelet, the error can be computed efficiently in an accumulative manner.展开更多
基金Supported by the Natural Science Foundation of China (50075032) and State High-Technology Development Program of China (2001AA421150)
文摘A strategy for B-spline curve data reduction based on non-uniform B-spline wavelet decomposition is presented. In existing methods of knot removal, ranking the removal knots depends on a procedure of assigning a weight to each knot to indicate its significance. This is reasonable but not straightforward. Propose is a more straightforward and accurate method to calculate the weight. The wavelet coefficient is taken as a weight for the corresponding knot. The approximating curve and the error can be obtained directly from the wavelet decomposition. By using the hierarchical structure of the wavelet, the error can be computed efficiently in an accumulative manner.
