Wide-Angle Augmentation for SAR Imaging via Generative Spatial-Frequency Models

In this work we present a K-space augmentation algorithm based on the use of deep generative models to improve cross-range resolution for monostatic synthetic aperture radar (SAR). The approach is based on the use of weighted adversarial training in the spatial-frequency domain. Initial efficacy of the approach is established through computational electromagnetic simulations of conducting targets in 2D.