代数重建法(ART)的图像重建质量和重建时间受到许多因素的影响。采用计算机仿真试验, 研究了先验知识、初始值、不同噪声下不同投影数、缺角度等对图像质量及松弛参数选取的影响。试验结果表明, 适当地选取初始值及加入先验知识可以极大地提高图像质量, 同时, 先验知识的加入影响着最佳松弛参数的选取; 不同噪声下不同投影数对松弛参数的选取也不同; 在缺失一定角度投影的情况下, 可以通过均匀地增加投影数来提高图像质量。

Quality and time of the image reconstruction with algebraic reconstruction techniques(ART) algorithm is influenced by many factors. The influences of some factors including the prior knowledge, the initial value, different projection numbers under different noise levels and projection data with limited angles on quality of the image reconstruction and selection of relaxation parameter, were investigated with computer simulation experiments. The experiment leads up to the result that selecting the proper relaxation parameter and adding the priori knowledge can greatly improve the image quality. At the same time, the priori knowledge and different projection numbers under different noise levels also impact the selection of optimum relaxation parameter. Under the circumstance of projection data with the lack of a certain range of observation angles, the quality of image can be improved by evenly increasing the projection numbers.