Electrical capacitance tomography image reconstruction based on singular value decomposition

Abstract

Electrical capacitance sensor is a soft-field sensor; the number of possible independent measurements is very small compared to the number of pixels required for an acceptable image. These problems make it difficult to reconstruct good quality images on-line. Although images reconstructed using the linear back-projection (LBP) algorithm are generally qualitative rather than quantitative, and various iterative methods have been developed to obtain improved image quality, the LBP algorithm is still the most commonly used algorithm because it is fast and hence suitable for on-line use. A new reconstruction method based on singular value decomposition of re-normalized sensitivity matrix is presented. Simulation research shows that it has reconstruction quality similar to that of an iteration method, and reconstruction time similar to that of the LBP algorithm.