On the effect of support estimation and of a new model in 2-D inverse scattering problems

Starting from the concept of "degree of non-linearity" of the relationship amongst unknown parameters and scattered fields, we discuss how knowledge of the geometrical features of the targets favorably affects the solution of inverse scattering problems. The same concept is also exploited to optimize the contrast source-extended Born (CS-EB) model, resulting in the optimized CS-EB (O-CS-EB) inversion approach, which remarkably reduces false solutions occurrence. Then, a two-steps approach is proposed as a reliable and effective inversion strategy. The first step consists in retrieving the shape of the targets using the linear sampling method, while in the second step, the estimated shapes and the O-CS-EB are exploited to reconstruct the permittivity profile. The approach has been tested against experimental data, showing that very accurate reconstructions are achieved from monochromatic data.