Two Dimensional Electrical Impedance Tomography Research Based On Finite Element Method

Electrical impedance tomography(EIT) technology is a new kind of nondestructive medical detection and imaging technology. Image target is internal resistance distribution or change. Compared with conventional detection method,EIT is innocuous to human’s bodies.So it can be used repeatedly. And it is cheap and easy to carry.And it can detect the changes of the biological function.So it has great significance in terms of disease prevention, diagnosis and treatment,which has a widespread prospect.The paper introduces the principle of electrical resistance tomography technology and makes a model based on two dimensional electrical impedance tomography.The finite element method is used to solve the forward problem,derive the variational form of Laplace equation,and set up the finite element equation by linear interpolation. Based on Matlab software,we split the circular field and realize the denser partition of the border area.Gauss elimination method is used to solve the finite element equation to obtain the potential information.Based on the results of the positive problem, inverse problem of electrical impedance tomography is studied, nonlinear, and ill conditioned. Based on the study of classical Newton iterative algorithm, use a modified Newton Raphson algorithm based on Tikhonov regularization to solve inverse problem. make reconstructed images of single target and multi-target.And then how the relative position of goals and different regularization parameters can affect reconstructed image quality.Because the accuracy of reconstructed electrical parameters is not high based on modified Newton – Raphson algorithm and is too dependent on the split model.And the iterative algorithm need more time and space.We use Support Vector Regression principle to solve the problem of electrical parameters reconstruction.The algorithm is based on the statistical learning theory and it can solve high dimension and nonlinear problems.It is found that this method can achieve relative accurate results rapidly.To improve the accuracy of the reconstructed electrical parameters,we optimizate parameters based on genetic algorithm.It is found that this merhod is feasible and effective.