复杂流场的代数迭代重建算法  

Simple Self-correlative Algebraic Reconstruction Technique for Complicated Flow Field

作　　者：李尊营[1] 宋一中[2] 

机构地区：[1]临沂师范学院物理系,山东临沂276005 [2]烟台大学光电学院,山东烟台264000

出　　处：《光电子．激光》2006年第10期1259-1263,共5页Journal of Optoelectronics·Laser

基　　金：山东省科技计划资助项目(J04A63)

摘　　要：提出了简单自相关代数迭代重建算法(SSART)。采用数值模拟的方法,对当前有代表性的基本代数迭代法(ART)、联合代数迭代法(SART)、改进的联合代数迭代法(MSART)以及SSART进行了计算模拟,对比分析了各种算法的模拟效果和重建精度。结果发现,SSART的均方误差(MSE)在10-4数量级,比ART的降低了78.2%;峰值相对误差(PE)在10-2,降低了92.9%;计算程序上,SSART仅比ART多1个除数因子。所以SSART具有算法简单、重建精度高、稳定性好和自愈能力强等优点。An improvement was brought forward on algebraic reconstruction technique （ART）, here, named simple self-correlative algebraic reconstruction technique（SSART）. With numerical simulations, a model was reconstructed by basic ART, simultaneous ART （SART） ,modified SART（MSART）. SSART was simulated at the same time. Simulated results and reconstructive accuracy of all kinds of ARTs mentioned above were contrastively discussed on mean-square error （MSE）, absolute value error （AVE）, and peak error （PE）. As a result,the reconstructive accuracy is improved much by SSART. The MSE of SSART is decreased about 78.2% which is smaller than that of ART at the level of 10^-4 magnitude,and PE is decreased 92.9% at the level of 10^-2 magnitude. SSART is considered as a very superior reconstruction technique that has many good virtues such as simple iteration program, high reconstructive accuracy, high stability, and strong self-reconvering ability.

关 键 词：简单自相关代数迭代法(SSART) 代数迭代法(ART) 联合代数迭代法(SART) 光学计算机 层析(OCT) 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]