基于差分迭代的电阻抗成像算法研究  

An Algorithm of Electrical Impedance Tomography Based on Differential Iteration

作　　者：章伟睿 张涛 史学涛 付峰 徐灿华 Zhang Weirui;Zhang Tao;Shi Xuetao;Fu Feng;Xu Canhua(Department of Biomedical Engineering,The Fourth Military Medical University,Xi’an 710032,China;Drug and Instrument Supervision and Inspection Station,Xining Joint Logistics Support Center,Lanzhou 730050,China)

机构地区：[1]空军军医大学军事生物医学工程学系,西安710032 [2]西宁联勤保障中心药品仪器监督检验站,兰州730050

出　　处：《电工技术学报》2021年第4期747-755,共9页Transactions of China Electrotechnical Society

基　　金：国家自然科学基金(31771073,61771475,51837011);军委科技委基础加强计划(2019-JCJQ-JJ-096)资助项目。

摘　　要：大多数电阻抗断层成像算法需要选择较优正则化参数来克服方程病态性,以获得较好的图像质量。该文提出一种基于差分迭代的电阻抗断层成像算法,利用动态线性逼近在不调整正则化参数的情况下提高成像质量。在近似线性变化区域内建立扰动模型,利用梯度法推导出电导率差值迭代关系进行快速重构成像,并将重构图像与基于几种客观正则化参数选取方法的重构图像在位置误差、分辨率、形状误差、环状伪影等性能指标方面进行比较。仿真结果证明,该文提出的差分迭代算法能与常用重构算法保持相似或更优的效果,该方法具有较好的实际应用前景。Most electrical impedance tomography algorithms need to select optimal regularization parameters to overcome the ill-conditioned equation and obtain better image quality.This paper proposes an electrical impedance tomography algorithm based on differential iteration,which uses dynamic linear approximation to improve imaging quality without adjusting the regularization parameters.A disturbance model is established in the approximate linear region,and the gradient method is used to derive the conductivity differential iterative relationship for rapid reconstruction.Then,the image reconstructed by this algorithm is compared with the reconstructed image based on several objective regularization parameter selection methods in terms of position error,resolution,shape deformation,ringing,and so on.The simulation results show that the proposed differential iterative algorithm can maintain similar or better results with the commonly used reconstruction algorithms,and has good practical application prospects.

关 键 词：电阻抗断层成像差分迭代 多目标成像 快速重构算法 正则化参数选取 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术] R318[自动化与计算机技术—计算机科学与技术]