A sparsity reconstruction algorithm for electrical capacitance tomography based on modified Landweber iteration

To improve the image quality and reduce the computational cost for electrical capacitance tomography (ECT), a sparsity reconstruction algorithm based on a modified Landweber iteration is presented in this paper. A soft-thresholding function is constructed for permittivity distributions with a continuous phase of low and high permittivity, respectively. An optimal step length is used during the iterative image reconstruction process. The effect of the regularization parameter, the noise level in data, and the permittivity distribution on the performance of the proposed algorithm is discussed according to the correlation coefficient. The modified Landweber iteration with a zero regularization parameter is also implemented for comparison. Simulation and experimental studies were carried out and the corresponding computational costs were estimated, showing the sparsity reconstruction outperforms the Landweber iteration.