Research On Regularization Method For Image Reconstruction In Electrical Impedance Tomography

Electrical impedance tomography(EIT)is a visual measurement technique aimed to reconstruct the complex conductivity distribution in the detected region.Due to the advantages of noninvasion,radiation-free,fast imaging and low cost,it is widely applied in industrial process imaging,nondestructive testing of materials,geophysical exploration and biomedical imaging.However,the image reconstruction in EIT is an ill-posed inverse problem,which seriously hinders the development of the EIT technique.Firstly,to solve the problem that the smooth area of the reconstructed image is prone to staircase effect,a total generalized variational regularization method that can establish a good compromise between retaining edge information and suppressing the staircase effect is proposed.The objective function of the proposed method is solved by the primal dual algorithm,and validated with circular model in the simulation and experiment.In addition,in order to investigate the feasibility of the proposed method in medical imaging,the lung pathological model is simulated and investigated.Secondly,to improve the poor edge preserving performance of the hybrid total variation method,a non-convex hybrid total variation regularization method based on non-convex function is proposed.The regularization parameter,weighting factor and non-convex parameter are all selected by adaptive method.The proposed method is solved by alternating direction multiplier method and validated by models with single conductivity and hybrid conductivity in the simulation and experiment.The results show that the total generalized variational regularization method is effective in suppressing the staircase effect,and the non-convex hybrid regularization method shows the advantage of edge information retention.