Tomographic Sub-Surface Detection of Defects by means of a GA-based Inverse Scattering Procedure

In this paper, an assessment of the integrated genetic-algorithm strategy based on a numerically computed Green's function for subsurface inverse scattering problems arising in NDE/NDT industrial applications is presented. To show the effectiveness and current limitations of the approach in dealing with various scenarios characterized by lossless and lossy host media as well as in noisy environments, several numerical experiments are considered. The obtained results confirm the effectiveness of the approach in fully exploiting the available a-priori information through a suitable scattering model, which allows a non-negligible enhancement of the reconstruction accuracy as well as a reduction of the overall computational burden with respect to standard inverse scattering approaches.