Microwave Imaging in a Dielectric Half-Space Medium Bounded by a Metal Surface

In recent years, microwave imaging technology has been successfully used for nondestructive testing of cryogenic fuel tanks foam insulation in the space industry. In this application, solving the inverse scattering problem becomes complicated due to the presence of the metal surface of the tank that is a perfect reflector for the electromagnetic waves. One of the possible solutions is the well known Gauss-Newton iterative method using an appropriate Green's function and Tikhonov regularization. For the infinite half-space (with no metal surface) the unknown depth of a subsurface target can be found even for a one-frequency signal by the focus plane determination method that determines the focusing distance corresponding to the minimum of the integrated amplitude modulus of a reconstructed image. However, when the metal surface is present, for a mono-static configuration of the antenna system used in the laboratory experiments, the reconstructed image of a target is well focused at different depths located at a periodic interval, and it is impossible to detect the true depth by conventional methods. In a multi-static case with separated transmitting and receiving antennas, the target depth can be detected, but this requires corresponding equipment for obtaining the synthetic aperture. Another way to address this problem is using the multi-frequency signal. In this work, the results of reconstruction from both simulated and experimental data are presented.