Image reconstruction for ECT system based on combination electrodes

Electrical capacitance tomography sensor is a soft field sensor providing very small number of projection data, which makes it difficult to reconstruct good quality images on line. A new image reconstruction algorithm based on multiple linear regression and matrix singular value decomposition is proposed. Image reconstruction for 16-electrode sensor in combination electrode mode, that is, each detection and source electrode consists of two adjacent electrodes, is investigated. The projection data and the sensitivity distributions for non-combined 8-electrode and combined 16-electrode sensor are obtained using finite element analysis. Popular linear back projection (LBP), the algorithm based on Laridweber's iteration and the new algorithm proposed in this paper are used to reconstruct various permittivity distributions. Comparisons in reconstruction quality and in reconstruction time are made. Results show that combined 16-electrode sensor can produce better image quality than can non-combined 8-electrode sensor. The new algorithm proposed in this paper has reconstruction quality similar to that of the algorithm based on Landweber's iteration, and reconstruction time similar to that of the LBP algorithm.