Comparative analysis of performance for optimization-based transionospheric SAR autofocus

Abstract

For a spaceborne synthetic aperture radar (SAR) operating on low frequencies (such as P-band), turbulence in the Earth's ionosphere may cause significant phase perturbations of the interrogating signals. These perturbations depend on both the antenna and target coordinates and may lead to substantial image distortions. In our previous work, we proposed a variational approach to correcting the distortions that we called the transionospheric SAR autofocus. It required solving a complex optimization problem but performed well in numerical tests. As the optimization problem may be considered a hurdle, in the current work we compare the performance of the transionospheric SAR autofocus against that of a non-variational approach. The latter combines partial focusing with traditional autofocus where the phase perturbations are assumed to depend only on the antenna coordinates but not the target coordinates. In most cases, the optimization-based SAR autofocus produces images with better articulated peaks (i.e., peaks that are taller and narrower) as compared to those by the alternative approach.