基于GMRES和Tikhonov正则化的生物电阻抗图像重建算法  

A Hybrid Reconstruction Method in Electrical Impedance Tomography Based on GMRES and Tikhonov Regularization

作　　者：王化祥[1] 范文茹[1] 胡理[1] 

机构地区：[1]天津大学电气与自动化工程学院,天津300072

出　　处：《生物医学工程学杂志》2009年第4期701-705,共5页Journal of Biomedical Engineering

基　　金：国家自然科学基金资助项目(60532020;50337020;60472077;60301008);国家科技部支撑计划项目资助(2006BAI03A00)

摘　　要：电阻抗层析成像(Electrical impedance tomography,EIT)是利用被测物体场内部电导率分布不均匀性,通过边界注入电流,测量边界电压变化,重构被测场内电导率分布图像。由于EIT测量数据有限,场域存在严重的非线性,导致问题的欠定性。我们介绍了一种新的组合算法,利用GMRES算法生成Krylov子空间,并结合Tik-honov正则化方法进行图像重建。该算法不仅改善了实时性,而且提高了成像质量及鲁棒性。Electrical impedance tomography （EIT） is a technique for reconstructing the conductivity distribution of measured field owing to its characteristics of being non-homogeneous, of injecting current at the boundary of the measured subject,of measuring the corresponding changes in voltage, and of reconstructing the image of the subject consequently. However, the limited measurement data of EIT, and the serious nonlinearity of the field result in illposed problem, and the resolution of reconstructed image is poor. To solve the problem, a new hybrid algorithm is herein proposed. The method combines the characteristics of Krylov subspace and Tikhonov regularization, which can improve the real time performance, the quality and robustness of reconstructed image.

关 键 词：电阻抗层析成像 TIKHONOV正则化 KRYLOV子空间 组合算法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]