Spaceborne Synthetic Aperture Radar (SAR) is an invaluable tool for Earth Observation (EO), providing high-resolution, all-weather, day-and-night imaging capabilities essential for continuous environmental monitoring and structural deformation tracking. Within this context, spaceborne SAR faces a trade-off between wide-area spatial coverage and azimuth resolution. Wide-swath acquisition modes actively steer the beam, truncating the integrated Doppler bandwidth, leading to a reduced azimuth resolution in the focused imagery. Efforts to improve this resolution, labelled Azimuth Super-Resolution (ASR), attempt to invert this degradation. However, classical methods often suffer from rigid signal priors, whereas standard deep learning architectures suffer from regression-to-the-mean, which smooths out high-frequency details and often are limited to amplitude-only. Consequently, this work introduces DiffASR: a conditional Denoising Diffusion Probabilistic Model (DDPM) designed for complex-valued SAR ASR.

The proposed methodology develops a hybrid, timestep-aware loss function that incorporates the structural and phase performance of the reconstruction in the pixel space. Simulating the truncated Doppler bandwidth is achieved through a forward subaperture decomposition model with varying broad-side frequency bands, progressively retaining 87.5%, 75%, 62.5%, or 50% of the original acquisition bandwidth. The model is trained on these discrete bands and evaluated on structural and radiometric metrics, alongside isolated point target and phase analysis.

The results demonstrate that while DiffASR shows a strong capacity to successfully sharpen the azimuth spatial resolution mainlobe and reconstruct native backscatter statistics, the stochastic generation introduces notable penalties in phase decorrelation and sidelobe energy dispersion. Ultimately, we establish DiffASR as an effective candidate for the super-resolution of point targets in amplitude-only applications, but conclude it can only be implemented for phase-dependent downstream tasks, such as Interferometric SAR (InSAR), with operational limitations. ... Read more

In the last few decades, a large increase in interest in space and particularly the Moon has taken place. The Moon is seen as a gateway to the rest of the Solar System. Missions to the Moon will inevitably lead to technological and scientific advancements. These would help in humanity’s mission to explore and develop habitats in the Solar System. Companies see economic opportunities in these places for activities such as the acquisition of rare Earth materials, as well as commercialising space travel. Furthermore, countries see these accomplishments
as a sort of international competition while also collaborating with other nations. The mission design presented here aims to facilitate these objectives by providing the necessary navigation support to any future mission on or around the Moon... ... Read more