Two-Stage Spatial-Frequency Joint Learning for Large-Factor Remote Sensing Image Super-Resolution

Super-resolution (SR) neural networks have recently achieved great progress in restoring high-quality remote sensing images (RSIs) at low zoom-in magnitudes. However, these networks often struggle with challenges like shape distortion and blurring effects due to the severe absence of structure and texture details in large-factor remote sensing image super-resolution (RSISR). Addressing these challenges, we propose a novel two-stage spatial-frequency joint learning network (TSFNet). TSFNet innovatively merges insights from both spatial and frequency domains, enabling a progressive refinement of SR results from coarse to fine. Specifically, different from existing frequency feature extraction approaches, we design a novel amplitude-guided-phase adaptive filter (AGPF) module to explicitly disentangle and sequentially recover both the global common image degradation and specific structural degradation in the frequency domain. In addition, we introduce the cross-stage feature fusion design to enhance feature representation and selectively propagate useful information from stage one to stage two. Quantitative and qualitative experimental results demonstrate that our proposed method surpasses state-of-the-art techniques in large-factor RSISR. Our code is available at https://github.com/likakakaka/TSFNet_RSISR.