Image reconstruction for ECT based on high-order approximate sensitivity matrix

Abstract

Electrical capacitance tomography (ECT), widely used for industrial measurements, reflects the permittivity distribution of the imaging region by measuring capacitance. The image reconstruction method is a crucial factor in determining image quality. Normally, the ECT image is reconstructed by sensitivity matrix. To alleviate the effects of the soft field property of the traditional sensitivity matrix, we propose an ECT reconstruction method based on the high-order approximate sensitivity matrix. The high-order approximate sensitivity matrices with four different nonlinearities, including a traditional sensitivity matrix, are derived through electromagnetic field analysis. Both numerical simulation and experiment are implemented for four sensitivity matrices using the Tikhonov regularisation algorithm and Landweber algorithm. Sensitivity evaluation indicators, image evaluation metricises, and capacitance matching correlation coefficients are used to compare the performance of sensitivity matrices. The results show that the high-order approximate sensitivity matrix with suitable nonlinearities improves image reconstruction quality. In different flow regimes, the sensitivity matrix with proper linearity has obvious advantages for regimes with high nonlinearity. The high-order approximate sensitivity matrix leads to shorter calculation time and can be used by various algorithms to enhance image reconstruction quality.