A compressive sensing approach to NDE/NDT

This paper presents an innovative microwave imaging technique for detecting defects in dielectric or conductive media developed through the Bayesian Compressive Sensing theory. Thanks to the a-priori knowledge of the medium unaffected by defects, the inversion problem is formulated as the estimation of the sparse differential equivalent currents within the support of the defects. The strategy has been developed in a multi-task implementation in order to efficiently handle the information provided by the scattered data.