基于改进SIRT算法的声学CT温度场重建  

Temperature field reconstruction of acoustic CT based on improved SIRT algorithm

作　　者：颜华[1] 孙灿烽 王伊凡[1] YAN Hua;SUN Can-feng;WANG Yi-fan(School of Information Science and Engineering,Shenyang University of Technology,Shenyang 110870,China)

机构地区：[1]沈阳工业大学信息科学与工程学院,沈阳110870

出　　处：《沈阳工业大学学报》2021年第5期550-556,共7页Journal of Shenyang University of Technology

基　　金：国家自然科学基金项目(61372154,61071141);辽宁省博士启动基金项目(201601157).

摘　　要：针对当重建区域划分的网格数多于声波路径数时,重建温度场中通常会出现很多噪点的问题,提出了一种改进SIRT算法,将平滑滤波器融入SIRT算法,通过专门设计的内、外平滑滤波器,抑制重建温度场中的噪点.为验证所提方法的有效性,对3个典型的模型温度场和一个实际的单峰温度场进行了重建.结果表明,与常用的最小二乘法、ART算法和SIRT算法相比,改进SIRT算法的重建误差明显降低.改进的SIRT算法具有更好的复杂温度场重建能力.Aiming at the problem that a lot of noise points usually appear in the reconstructed temperature field when the number of grid cells in the reconstruction area is larger than the number of acoustic paths,an improved SIRT algorithm with the integration of smoothing filter was proposed.By using specially designed inner and outer smoothing filters,the noise points in the reconstructed temperature field were suppressed.In order to verify the effectiveness of as-proposed method,three typical model temperature fields and an actual single-peak temperature field were reconstructed.The results show that the reconstruction errors of improved SIRT algorithm can be obviously reduced in comparison with the least square method(LSM),ART algorithm and SIRT algorithm.The improved SIRT algorithm has better reconstruction ability for complex temperature fields.

关 键 词：声学CT 温度场重建 重建算法 SIRT算法 平滑滤波器 重建误差 最小二乘法 ART算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]