Study On Computational Imaging Method In Electrical Capacitance Tomography

Electrical Capacitance Tomography(ECT)technology as a powerful visualization tomography technique has been widely used in different industry scenarios.However,the application of ECT is plagued by low-quality reconstructions.To address the challenge,this paper mainly studies the computational imaging method of ECT to improve the image quality,and the main work can be summarized as follows:(1)The mathematical principles,advantages and disadvantages of conventional iterative and non-iterative imaging algorithms are qualitatively analyzed,and the numerical performances of the linear back projection algorithm,Tikhonov regularization algorithm,Land weber iterative algorithm,conjugate gradient algorithm,algebraic reconstruction algorithm and simultaneous iterative reconstruction technique are evaluated quantitatively.(2)A new cost function based on sparsity prior of imaging targets is proposed to improve the reconstruction quality,which can transform the ill-posed ECT image reconstruction problem into an optimization problem,and the two-step iterative shrinkage/thresholding method is used to solve the built cost function efficiency.Experimental results indicate that the proposed imaging method can ensure the sparsity and stability of the solution,reducing artifacts,reconstructing the details of imaging targets.(3)In order to improve the reconstruction quality,a powerful imaging model is proposed to emphasize the inaccuracy of the measurement data and model deviation,and a new cost function is proposed to model the ECT inverse problem,in which the L1 norm is used as the data fidelity term to alleviate the sensitivity of solution to noises and the L1 norm of the unknown variable is used as regularizer to strengthen the assumed sparsity of the imaging targets.The soft thresholding method and the fast-iterative shrinkage thresholding technique are embedded into the iterative split Bregman framework to generate an effective numerical method to solve the devised cost function.The numerical results indicate that the proposed ECT imaging technique not only leads to the improved imaging quality but also improves the robustness.(4)In order to reduce reconstruction errors and artifacts,a novel cost function is built for imaging,in which the L1-2 norm is designed as a regularizer for encoding the sparsity prior of imaging objects.The ant lion optimizer algorithm,the the alternating direction method of multipliers are combined into the memetic algorithm as a powerful optimizer for solving the built cost function more effectively.The numerical results indicate that the proposed imaging technique has higher accuracy compared with other popular imaging techniques,improving the quality of reconstructed images and the global convergence.(5)To weaken the impact of measurement noise on image quality,a denoising method based on the variational mode decomposition is put forward,which can reduce the difference between collected capacitance values and real capacitance values.The proposed method uses the variational mode decomposition method to extract the mode components in capacitance signal,designs a more targeted filtering method,and uses the thresholding filtering method optimized by the manta rays foraging optimization algorithm,reducing the noise of each mode component.The numerical results indicate that the proposed method can mitigate the noise in capacitance signal effectively,improving the measurement quality of capacitance data,and laying a foundation for the reconstruction of high-quality images.