The regularization method is commonly used for image optimization in ECT.The L1 norm is overly neglecting the image features and the L2 norm is too smooth,which leads to poor imaging quality,an ECT regularization image reconstruction method based on the L1+p norms is proposed.Using the L1+p(0<p<1)paradigm as the regularization term of the loss function,the ECT image reconstruction is made both sparse and smooth by adjusting the value of the parameter p.The simulation experiment's results showed that the correlation coefficient of the improved regularization algorithm with norm is on average improved by 77.04%compared with the LBP algorithm,by 36.18%compared with Tikhonov regularization algorithm and by 41%compared L1-norm regularization algorithm in complex multi-bubble flow imaging.Gassolid two-phase flow experiments showed that the proposed algorithm can effectively improve the imaging quality,and is capable of real time adjusting image sparsity.
万方数据
