CoRE-UIR: Prior-guided common and residual experts for efficient all-in-one remote sensing image restoration

2026-07-30

AI Summary

This work addresses the challenge of composite degradation in remote sensing imagery, which induces both global distribution shifts and local structural distortions. Existing unified restoration methods often suffer from weak degradation cues and excessive parameter redundancy. To overcome these limitations, the authors propose CoRE-UIR, a novel framework that decouples expert modules into dense generic experts and low-rank residual experts. The approach further incorporates explicit degradation prior embedding (DPE) derived from CLIP and global feature modulation (GFM) to enable efficient, adaptive restoration. Additionally, the study introduces MDVD-108K, the first large-scale UAV image restoration dataset encompassing diverse composite degradations. Experiments demonstrate that CoRE-UIR achieves an average PSNR gain of 1.05 dB, 11.83× faster inference, and 85.3% lower peak memory consumption across multiple benchmarks, while exhibiting strong generalization in downstream tasks and unseen degradation scenarios.

Abstract

Remote sensing images acquired by unmanned aerial vehicles (UAVs) and satellites are often degraded by adverse weather, illumination variation, and imaging artifacts, which may co-occur and jointly induce global distribution shifts and local structural corruption. Although All-in-One image restoration offers an appealing unified alternative to task-specific pipelines, existing methods still suffer from weak or implicit degradation cues and parameter redundancy caused by full-rank multi-expert designs with overlapping restoration behaviors. We propose CoRE-UIR (Common and Residual Experts for Universal Image Restoration), a prior-guided global-local framework centered on the Common-and-Residual Expert Block (CoRE). CoRE explicitly decomposes restoration capacity into a common dense expert for degradation-invariant restoration and low-rank residual experts for degradation-specific compensation, enabling adaptive specialization without redundant expert replication. Built on this design, Degradation Prior Embedding (DPE) adapts frozen CLIP features into an explicit restoration-oriented prior, while Global Feature Modulation (GFM) aligns global feature statistics before local residual compensation. We also construct MDVD-108K (Multi-Degradation VisDrone), a large-scale UAV restoration dataset covering both single and compound degradations, together with a real-world test set. Extensive experiments on multiple datasets show that CoRE-UIR improves the overall average PSNR by 1.05 dB while running 11.83$\times$ faster and reducing peak memory by 85.3% relative to the strongest baseline, BaryIR, thereby maintaining a favorable quality-efficiency trade-off. Evaluations on downstream tasks and unseen degradation also validate the generalizability of CoRE-UIR. The code and dataset will be released at https://github.com/zzaiyan/CoRE-UIR.

Problem

• remote sensing image restoration
• all-in-one restoration
• image degradation
• parameter redundancy
• multi-degradation

Innovation

• Common-and-Residual Expert
• Degradation Prior Embedding
• Global Feature Modulation
• All-in-One Restoration
• Low-Rank Experts