Improved electrical capacitance tomography algorithm based on homotopy perturbation regularization.

Electrical Capacitance Tomography (ECT) is one of typical tomography technologies based on capacitance-sensitive fields, which can reconstruct the distribution images of permittivity distribution of different fluid in the measured field and is often used for realizing industrial detection in specific occasions. Landweber and homotopy perturbation regularization algorithms are used for image reconstruction in ECT system, but these algorithms are with many iteration steps, slow convergence speed, and relatively low quality of the reconstructed images. Aiming at these problems, this paper proposes an improved image reconstruction algorithm. Firstly, a regularization term is added to the target function by using the different dielectric constant distribution between the current moment and the previous moment. Then, using the homotopy perturbation to derive the second-order iterative formula to get the homotopy perturbation regularization algorithm, and finally the improved algorithm is obtained by combining the landweber algorithm based on weight factor. Furthermore, the improved homotopy perturbation regularization algorithm is applied for ECT image reconstruction. The numerical simulation experiment results show that the improved algorithm is with the highest comprehensive scores for the four flow patterns of laminar flow, annular flow, core flow and bubble flow, which is higher than the landweber algorithm and the improved homotopy perturbation regularization algorithm by 3% ~ 20%, and 18% ~ 31%, respectively. The relative error, correlation coefficient and subjective effect of the reconstructed image are significantly improved, and it has certain anti-noise performance. These reflect the improved algorithm is with practical value. [ABSTRACT FROM AUTHOR]