Hybrid iterative reconstruction method for imaging problems in ECT

Abstract

As a powerful visualization tomography technique, electrical capacitance tomography (ECT) method, which reconstructs the permittivity distribution in a sensing domain according to the measured capacitance data, has been widely used in different industry scenarios. However, the image reconstruction in ECT is an ill-posed problem, and its application is plagued by low-quality reconstructions, and seeking for a robust algorithm to reduce reconstruction errors and artifacts is crucial. To increase the reconstruction quality, a novel cost function is built for imaging, in which the $L_{1-2}$ norm is designed as a regularizer for encoding the sparsity prior of imaging objects. The ant lion optimizer (ALO) algorithm and the alternating direction method of multipliers (ADMMs) with the Douglas–Rachford splitting (DRS) method and the soft thresholding algorithm as powerful optimizers for minimizing the subproblems are combined into a novel solver for solving the built cost function more effectively. Simulation and experimental results indicate that the proposed imaging technique can improve the quality of reconstructed images.