SR-CT技术乘型迭代算法的研究

The Research on Multiplication Algebraic Arithmetic of SR-CT Technique

算法为SR-CT(synchrotron radiation computed tomography)技术中的重要组成部分,其中乘型迭代算法为该技术的一种有效算法。本文对乘型迭代算法进行了研究,并对该算法中的一些重要参数,如迭代步长、迭代次数及初始解进行了优化分析,给出了这些参数对重建图像及运算时间的影响关系以及这些参数间的相互影响关系,同时给出了获取较优初始解的方法以及迭代步长和迭代次数的最佳取值范围。

In the technique of synchrotron radiation computed tomography （SR-CT）, the arithmetic is an important part. And the multiplication algebraic arithmetic is an effective arithmetic of SR-CT technique. In order to understand the relation between reconstruction images and the multiplication algebraic arithmetic, the arithmetic is studied in this paper. The important parameters in the arithmetic such as size of iterative step, iterative times and initial solution are analyzed and optimized in detail. As a result, the relations of reconstruction images quality and computational time affected by the three parameters above are acquired. The impact models among these parameters are also obtained. The method to get the better initial solution is given. And the ranges of selecting the iterative step and iterative times are gained.