ECT Imaging Based on Fractional-Order Particle Filtering and Image Super-Resolution

Electrical Capacitance Tomography (ECT), as a rapidly developing process tomography technique, has been widely applied in multiphase flow monitoring, industrial process control, and medical imaging. However, due to the “soft-field” effect, the ill-posed nature of the inverse problem, and limitations in data acquisition. Existing reconstruction methods generally suffer from large reconstruction errors, low accuracy, blurred edges, and pronounced artifacts. To address these issues, this paper proposed a reconstruction method based on fractional-order particle filtering and image super-resolution. Firstly, a fractional-order formulation is introduced into the system equation to better characterize the dynamic behavior of the ECT system, by exploiting the short-term memory property of fractional-order systems, thereby accelerating convergence in the search space. Subsequently, the particle filter is employed to effectively alleviate the “soft-field” effect of the ECT system. By directly approximating the posterior probability distribution of the system state through Monte Carlo sampling, more informative particles can be selected, which enhances reconstruction accuracy and stability. Finally, an image super-resolution model based on Lorentzian function fitting is adopted, where the point spread function is estimated and combined with the Lucy–Richardson algorithm to achieve high-resolution imaging and mitigate edge blurring.In the experimental study, a three-dimensional simulation model was established using COMSOL, and a 12-electrode hardware platform was constructed for validation under various typical flow patterns. The results demonstrate that the proposed method achieves superior image quality while maintaining low computational cost (less than 0.3 s), providing an effective solution for ECT imaging in complex industrial environments.