论文部分内容阅读
代数重组技术(简称ART)是级数展开算法这一大类中一种重要的方法。ART的计算效率是比较低的,但是,在某些情况下,比如折射修正、非线性修正或不完全投影重组,ART可以优于卷积法。在本文中,对ART算法的基本步骤进行了系统的计算机模拟研究,例如迭代方法、迭代步骤、加权因子的选择等。此外,对弛豫系数的选择进行了讨论,为改进重组精度和计算效率作了一些尝试。最后编制了实用的程序并给出了使用这程序所得到的图象。
Algebraic recombination technology (referred to as ART) is an important method in the category of series expansion algorithm. The calculation efficiency of ART is relatively low, however, in some cases, such as refractive correction, nonlinear correction or incomplete projection recombination, ART can be better than convolution. In this paper, the basic steps of ART algorithm are systematically studied by computer simulation, such as iterative methods, iterative steps, selection of weighting factors and so on. In addition, the choice of relaxation coefficient is discussed, and some attempts are made to improve the recombination accuracy and computational efficiency. At last, a practical program is compiled and the image obtained by using this program is given.