Modified newton-raphson method using a region of interest in electrical impedance tomography

Abstract

Electrical impedance tomography (EIT) visualizes the distributions of the electrical characteristics of an unknown object by using reconstruction algorithms. In the EIT system, several electrodes are installed around the phantom’s periphery, and an especially-designed current is injected through the installed electrodes. The injected current induces voltages from the object’s boundary, and there induced voltages are used as an electrical characteristic distribution that visualizes the unknown object. However, EIT has an inverse problem, the ill-posedness in the Hessian matrix, which causes several problems. The problems caused by the ill-posedness, such as modeling errors in the linearization of the nonlinear measurement function and the noise included in the measured voltages, cause a bad performance in the image reconstruction. In this paper, we propose region-of-interest (ROI). This paper proposes ROI method for two-phase flow visualization, and we are modified Newton-Raphson (mNR) as an inverse solver to estimate the resistivity distribution inside a circular domain. Different weights are assigned to the object and the background regions, thus increasing the sensitivity and reducing the ill-posedness. Numerical simulations are carried out to validate the performance of the proposed method.