A New End-to-End Sar Super-Resolution Imaging Model

The resolution of SAR image plays a crucial role in representing the system's ability to distinguish two adjacent targets in space, making it an essential parameter for measuring imaging system performance. The necessity to enhance the resolution of SAR images after imaging has become imperative, primarily due to hardware constraints and the absence of comprehensive prior knowledge. In recent years, a surge in image super-reconstruction methodologies driven by deep learning has been witnessed. Given that SAR images are complex images with both amplitude and phase information, this paper adopts a novel end-to-end model. It divides the input of the network into two channels, one for the real part and the other for the imaginary part, preserving both the amplitude and phase information of the SAR image.