Autofocusing Technique Based on Generalized Multilayer Stolt Migration

A method is proposed for estimation of geometric information (GI) for obliquely layered geometry. The GI is estimated in the process of autofocusing (AF) of a ground-penetrating radar image. A novel AF technique is proposed based on the generalized multilayer Stolt migration algorithm. In the algorithm, the position and angle of the layer boundaries are determined using a Hough transform. At each layer, the AF metric is iteratively evaluated to estimate the relative permittivity (RP) of the layer. The performances of four AF metrics are compared, and an RP determination algorithm is suggested to reduce the number of AF metric evaluations. The proposed algorithm is validated with numerical and experimental data.