An Iterative Reconstruction Algorithm for Electrical Capacitive Tomography Using Fuzzy System

Electrical capacitance tomography (ECT) is used to image the permittivity distribution of an object by measuring the mutual electrical capacitance between sets of electrodes mounted on its periphery. Reconstruction of the cross-section images in ECT system from the capacitance measurements is a nonlinear and illposed inverse problem. However, there are many algorithms to solve the inverse problem; they are not accurate in all cases because they based on a linear approximation. More accurate images could be obtained by using iterative techniques, although these techniques consume much time. The convergence of these iterative methods depends on the accuracy of the first image passed through the iterative loop. This paper discusses a new method to improve the images obtained from the iterative methods by making good estimation for the first image. The proposed method is based on Fuzzy Inference Systems (FIS) to predict the probability of the permittivity distribution in the first image. The first estimation by using fuzzy system is combined with iterative methods such as the iterative linear back projection (ILBP) technique to get more precise and fast images. The proposed method is tested for estimating the molten metal profile in lost foam casting (LFC) process which is one of the most energy efficient casting methods in the industry. The metal-fill profile is one of the important factors that affect casting quality. The proposed technique is able to detect the position of the metal by using just 8 measurements from the sensors. The reconstruction results from ECT data demonstrate the advantage of using the fuzzy system and show improvement of the images quality compared with the images obtained from the ILBP reconstruction method without using the fuzzy system.