Defect imaging based on closed-loop PID controlled Kalman Filter in electromagnetic tomography

Abstract

In this paper, a novel closed-loop PID controlled Kalman Filter is investigated for defect distribution reconstruction in electromagnetic tomography (EMT). The effectiveness of the Kalman Filter algorithm is heavily influenced by two key parameters: the observation noise covariance and the initial estimation error covariance. These parameters are crucial for the accuracy of the reconstruction process. In general, these two parameters are kept at a fixed value, and inappropriate choice of parameters will lead to artifacts and blurred boundary in defect reconstruction. In order to solve the above problem, a closed-loop PID controller is introduced to adjust the observation noise covariance and the estimation error covariance, thereby improving the defect reconstruction quality. Aiming at six different defect distributions, simulation and experiment are conducted to evaluate the reconstruction quality using the proposed method. Compared with the three classical reconstruction, the method proposed in this paper has clearer imaging and fewer artifacts, and the reconstruction correlation coefficient and relative reconstruction error are improved by 7.6 % and 23.3 % at least.