An Image Reconstruction For Electrical Capacitance Tomography Using Parametric Level Set

Abstract

Image reconstruction algorithm is essential for electrical capacitance tomography (ECT), which is still in the stage of popular research. With the development of image reconstruction algorithm, high-quality image is the key challenge for ECT all long. The paper proposes a kind of novel-image-reconstruction-algorithm for ECT using parametric level-set method to obtain high-image quality. Based on the relationship between dielectric constant distribution and capacitance value in the sensitivity area, parametric level set algorithm is capable of realizing absolute values ECT reconstruction. The paper presented simulation results of reconstructing the permittivity profiles of different water leakage using parametric level set method (PLS). Comparing with the state of the art image reconstruction algorithm, such as LBP regularization, landweber iterative algorithm and total variational regularization, the proposed method has better image quality, especially with high contrast multiphase data. PLS adopts Gaussian radial basis function (GRBF), which considerably reduces the number of unknowns. The parametric level set method can avoid the problem of regularization coefficients involved in the calculation process and reduce the Ill-posed Problem of image reconstruction. The proposed PLS method has demonstrated the superior image quality and better noise ratio (SNR).