Electronically Rotated Electrical Capacitance Tomography Sensing and Image Reconstruction Based on Transformer Model

Electrical capacitance tomography (ECT) is a noninvasive imaging technology based on capacitance sensors mainly used for measuring and imaging nonconductive fluid in multiphase flow. However, ECT has underdetermined and nonlinear problems due to limited sensor electrodes, which pose challenges for image reconstruction. Therefore, this paper uses the sensing technology based on electronically rotated electrical capacitance tomography (ER-ECT) to increase the number of independent measurements by combined excitation to alleviate the underdetermined problem. At the same time, ER-ECT combined with the Transformer model is used to reconstruct the dielectric constant distribution image from sensor measurement data, and the self-attention mechanism is used to solve the nonlinear mapping problem of ECT. The experimental results show that the proposed method is superior to the traditional method in terms of relative error and correlation coefficient and improves the quality and accuracy of image reconstruction of the sensor system.