Image Reconstruction of Capacitively Coupled Electrical Resistance Tomography Based on An Improved Kalman Filter Model

By developing an improved Kalman filter (KF) model, a new imaging method of capacitively coupled electrical resistance tomography (CCERT) is proposed for monitoring multiphase flows. In the improved KF model, the innovative idea of state evolution learning is proposed to learn the spatiotemporal correlations between the fluid states, and the level-based state representation is integrated to reduce the state dimension and improve the computation efficiency. Correspondingly, the imaging method based on the improved KF model is proposed for CCERT. Experimental results of the gas-liquid two-phase flow verify the effectiveness and potential of the proposed method. The improved KF model can quickly and effectively learn and utilize the flow distribution regularities across time frames, and can achieve better imaging performance in both quality and efficiency. The largest average relative image error (RIE) and the smallest average correlation coefficient (CC) of the images obtained by the proposed method are 0.125 and 0.831 for single-object distributions, respectively.