阻抗断层成像重构算法中的正则化方法

A regularized method in reconstructed algorithm of electrical impedance tomography

为降低电阻抗断层成像技术中 ,在有限元剖分规模较大时 ,重构算法的计算将出现严重病态性的问题 ,提出了一种修正重构矩阵 (Hessian矩阵 )的奇异值的正则化方法 ,在计算机模拟模型上进行了仿真重构研究 ,获得了较好的成像结果 ,并讨论了正则化参数对重构结果的影响 .

To solve the serious ill conditioned problem of reconstructed algorithm in electrical impedance tomography (EIT) when the FEM element number was bigger, we developed a new regularization method, which modified the singular value of reconstructed matrix. On the simulating computer model, we used it and obtained more satisfactory imaging results. The effect of regularization parameter on the results was discussed.