Complex-Valued SAR Image Super-Resolution via Subaperture Learning and Fusion

Synthetic aperture radar (SAR) imaging is a popular tool in many fields, while the contradiction between resolution and coverage is an open problem. The range resolution relies on the bandwidth of transmitted signals, while the azimuth resolution is dependent on the length of the synthesized aperture. They are related to radar configurations and are infeasible to be tuned freely. To solve the problem, a new SAR image super-resolution method is proposed in this article. It is composed of preimaging, feature learning, and reconstruction. In the first phase, the zero-padded technique is used to achieve spatial alignment. The result is then decomposed along the subapertures, forming the subaperture images and the subband images. They were delivered to the second phase for cross-domain learning and hierarchical fusion. The prelearning, encoder-decoder, and refinement are configured. The learned features are finally tuned for super-resolution reconstruction. On this basis, a new evaluation system was built to measure the performance. Both the supervised evaluation and the unsupervised evaluation are considered. An improvement of 3.78% is achieved for the amplitude components, while the phase accuracy is 4.62% better than the competitor.