Researchers have developed AstraMoE-SR, a novel single-image framework designed to simultaneously deblur satellite imagery and enhance its resolution. This method addresses the challenges posed by platform jitter in pushbroom satellite imaging, which causes spatially varying motion blur. Unlike existing approaches, AstraMoE-SR does not require auxiliary measurements or explicit blur kernel estimation. Instead, it infers the camera's movement by modeling degradation as local exposure trajectories and utilizes a conditional diffusion model to predict these trajectories, which then guide a latent diffusion backbone for alignment and reconstruction. The framework has demonstrated superior performance on the DOTA-v1.0 dataset, outperforming previous methods and a no-restoration baseline across multiple fidelity metrics. AI
IMPACT This research could improve the quality of satellite imagery for various applications by enabling more accurate deblurring and super-resolution without auxiliary data.
RANK_REASON The item is a research paper detailing a new AI model for image processing. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
