在计算机辅助几何设计(CAGD)领域,渐进迭代逼近(PIA)算法因其具有很好的自适应性和收敛稳定性,被广泛应用于插值与逼近问题.其中带权渐进迭代逼近(WPIA)算法通过调整向量加权明显加快了收敛速度.提出了一种带互异权值的渐进迭代逼近算法,不仅操作灵活,还可根据需要对各控制顶点进行调整,实现不同的迭代效果;同时通过引入一个参数,给出了可调权值迭代算法,当参数取合适值时,该算法的收敛速度比带权PIA算法更快,且权值取法不依赖于配置矩阵的特征值.最后用数值实例,通过对Bézier曲线、张量积Bézier曲面,以及三角Bézier曲面进行迭代,展示了该算法的有效性.
In CAGD, progressive iterative approximation (PIA) method is widely used to solve interpolation and approximation problems due to its perfect adaptability and convergence stability. Weighted progressive iterative approximation (WPIA) can accelerate the convergence rate by assigning an appropriate weight for each adjusting vectors. One new PIA method with mutually different weights is presented. It not only provides more flexibility in operation, but also achieves satisfactory iterative result for different control vertices. A set of weights with an ad- justable parameter has also been put forward, which can be obtained without resorting to the eigenvalue of colloca- tion matrices and can speed up the convergence rate compared with the WPIA method. Numerical examples of B6zier curves, tensor-product B~zier surfaces and triangular B6zier surfaces demonstrate the effectiveness of the method.