作者: Xianglong Liu , Ze Liu
DOI: 10.1016/J.FLOWMEASINST.2019.01.010
关键词: Regularization (mathematics) 、 Image resolution 、 Optimization problem 、 Algorithm 、 Iterative reconstruction 、 Inverse problem 、 Tomography 、 Tikhonov regularization 、 Computer science 、 Iterative method
摘要: Abstract Electromagnetic tomography (EMT) is a novel imaging modality of electrical tomography, which appears to be very promising. The precision and speed image reconstruction algorithms EMT are the keys its application in industrial biomedical fields. Image typical ill-posed ill-conditioned inverse problem. Following analyzing advantages disadvantages traditional Tikhonov regularization algorithm total variation algorithm, new objective functional introduced with L1 norm on data term Lp this paper, transforms problem into an optimization Besides, L1-Lp framework solved by approximation Gauss-Newton algorithm. Both numerical simulation experimental results demonstrate that proposed capable enhancing spatial resolution. It also proves has better performance terms patterns, compared such as linear back projection (LBP), standard projected Landweber iterative