Zero-Sharpen: A universal pansharpening method across satellites for reducing scale-variance gap via zero-shot variation

Hebaixu Wang, Hao Zhang, Xin Tian, Jiayi Ma

Abstract: Highlights•Zero-shot pansharpening (ZeroSharpen) can perform on a single pair of PAN/MS images.•Image-pairs-specific networks seek the non-linear spatial and spectral relationships.•Tailored loss functions for reducing the scale variance are employed for networks.•Novel variational model offers physical interpretability and stable convergence.•Numerous results prove the visual and statistical superiority of ZeroSharpen.