AnyBand-Diff: A Unified Remote Sensing Image Generation and Band Repair Framework with Spectral Priors

Existing diffusion models have made significant progress in generating realistic images. However, their direct adaptation to remote sensing imagery often disregards intrinsic physical laws. This oversight frequently leads to spectral distortion and radiometric inconsistency, severely limiting the scientific utility of generated data. To address this issue, this paper introduces AnyBand-Diff, a novel spectral-prior-guided diffusion framework tailored for robust spectral reconstruction. Specifically, we design a Masked Conditional Diffusion backbone integrated with a dual stochastic masking strategy, empowering the model to recover complete spectral information from arbitrary band subsets. Subsequently, to ensure radiometric fidelity, a Physics-Guided Sampling mechanism is proposed, leveraging gradients from a differentiable physical model to explicitly steer the denoising trajectory toward the manifold of physically plausible solutions. Furthermore, a Multi-Scale Physical Loss is formulated to enforce rigorous constraints across pixel, region, and global levels in a joint manner. Extensive experiments confirm the effectiveness of AnyBand-Diff in generating reliable imagery and achieving accurate spectral reconstruction, contributing to the advancement of physics-aware generative methods for Earth observation.

(1) Satellite images are essential for disaster response and environmental monitoring, but data is often missing due to clouds or sensor failures. Existing AI models fill in the gaps but ignore the physics of light, producing images that look real yet are scientifically unreliable. (2) We built AnyBand-Diff, an AI that learns the physical laws of how light reflects off the Earth. It can reconstruct missing spectral bands from any available subset—even when half the data is lost. (3) This makes AI-generated satellite data truly usable for tasks like calculating vegetation health or monitoring water bodies, supporting better real-world decisions.