🤖 AI Summary
Existing satellite video super-resolution (VSR) methods exhibit poor generalization on large-scale, complex scenes and struggle to model both long-term temporal dependencies and fine-grained local motion. To address this, we propose a novel temporal compensation paradigm based on local-global temporal differencing, introducing the first dual-module architecture—Short-Term Differencing Module (S-TDM) and Long-Term Differencing Module (L-TDM)—to separately capture inter-frame local motion and sequence-level global dependencies. We further design a Difference Compensation Unit (DCU) that leverages RGB inter-frame difference maps for guidance, employs bidirectional segment-wise differencing modulation, and incorporates spatio-temporal interaction to enhance feature consistency and temporal robustness. Our end-to-end differentiable VSR network achieves state-of-the-art performance across five mainstream satellite video benchmarks, with significant PSNR/SSIM improvements and superior perceptual quality. The code is publicly available.
📝 Abstract
Optical-flow-based and kernel-based approaches have been extensively explored for temporal compensation in satellite Video Super-Resolution (VSR). However, these techniques are less generalized in large-scale or complex scenarios, especially in satellite videos. In this paper, we propose to exploit the well-defined temporal difference for efficient and effective temporal compensation. To fully utilize the local and global temporal information within frames, we systematically modeled the short-term and long-term temporal discrepancies since we observe that these discrepancies offer distinct and mutually complementary properties. Specifically, we devise a Short-term Temporal Difference Module (S-TDM) to extract local motion representations from RGB difference maps between adjacent frames, which yields more clues for accurate texture representation. To explore the global dependency in the entire frame sequence, a Long-term Temporal Difference Module (L-TDM) is proposed, where the differences between forward and backward segments are incorporated and activated to guide the modulation of the temporal feature, leading to a holistic global compensation. Moreover, we further propose a Difference Compensation Unit (DCU) to enrich the interaction between the spatial distribution of the target frame and temporal compensated results, which helps maintain spatial consistency while refining the features to avoid misalignment. Rigorous objective and subjective evaluations conducted across five mainstream video satellites demonstrate that our method performs favorably against state-of-the-art approaches. Code will be available at https://github.com/XY-boy/LGTD.
