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

2026-05-14

AI Summary

This work addresses the spectral distortion and radiometric inconsistency commonly arising in remote sensing image generation with existing diffusion models, which often neglect underlying physical laws and thereby compromise the scientific utility of synthetic data. To overcome this limitation, the authors propose AnyBand-Diff, a novel framework that integrates gradients from differentiable physical models into the diffusion process for the first time. By leveraging physics-guided sampling to steer the denoising trajectory and combining mask-conditioned diffusion with a dual stochastic masking strategy, the method enables high-fidelity reconstruction of full spectra from arbitrary subsets of spectral bands. Furthermore, a multi-scale physics-aware loss function enforces spectral consistency jointly at pixel, regional, and global levels. Experiments demonstrate that AnyBand-Diff achieves physically plausible and spectrally accurate generation and band inpainting, substantially enhancing the scientific validity of synthesized remote sensing imagery.

Abstract

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.

Problem

• spectral distortion
• radiometric inconsistency
• remote sensing image generation
• physical laws
• spectral reconstruction

Innovation

• spectral prior
• physics-guided diffusion
• masked conditional diffusion
• radiometric consistency
• multi-scale physical loss