Compressive Processing of Microwave Data for NDE/NDT

摘要

In the framework of nondestructive evaluation and testing (NDE/NDT), this paper deals with the retrieval of unknown defects inside known dielectric media. A Bayesian Compressive Sensing contrast source inversion method enabled by a suitable definition of the inhomogeneous Green’s function is employed to fully exploit the a-priori information of the problem at hand, regarding the knowledge on the unperturbed structure and the physical sparseness of the perturbations. Selected numerical results are presented in order to assess effectiveness and robustness of the proposed methodology.