Interpretation-Oriented Cloud Removal via Observation-Anchored Residual Flow with Geo-Contextual Alignment

2026-07-02

AI Summary

Existing cloud removal methods prioritize visual realism at the expense of semantic consistency, often degrading performance in downstream remote sensing interpretation tasks such as semantic segmentation and change detection. To address this limitation, this work proposes the Geo-Anchored Cloud Removal (GACR) framework, which formulates cloud removal as an observation-anchored residual inverse problem. GACR introduces an Observation-Anchored Residual Flow (OAR-Flow) for physically interpretable and efficient reconstruction, and incorporates a Geospatial Context Prior Alignment (GCPA) mechanism to constrain the generation process within the semantic manifold guided by vision foundation models. This dual strategy effectively preserves both visual fidelity and spatial semantic integrity. Extensive experiments demonstrate that GACR consistently achieves superior reconstruction quality and significantly improves accuracy across six cloud removal benchmarks and twelve downstream interpretation tasks.

Abstract

Cloud removal (CR) is essential for optical remote sensing, serving as a prerequisite for reliable downstream interpretation, such as semantic segmentation and change detection. However, existing CR approaches often prioritize visual realism while overlooking their impact on subsequent analytical tasks, leading to semantic drift and degraded downstream performance. To address this issue, we propose Geo-Anchored Cloud Removal (GACR), a unified framework that jointly ensures faithful reconstruction and robust interpretability. At its core, GACR incorporates Observation-Anchored Residual Flow (OAR-Flow), which reformulates CR as a physically grounded residual inversion process. By anchoring the generative trajectory to the cloudy observation rather than pure noise, OAR-Flow enables fast, stable, and faithful reconstruction. To further preserve semantic structures critical for downstream interpretation, GACR integrates Geo-Contextual Prior Alignment (GCPA) to constrain the reconstruction within a semantic manifold induced by a Vision Foundation Model (VFM). Consequently, GACR strictly maintains the spatial-semantic integrity of complex landscapes. Extensive experiments across six CR datasets and twelve downstream tasks demonstrate that GACR produces superior reconstruction quality while consistently improving downstream task accuracy. The code is available at https://github.com/wzy6055/GACR.

Problem

• Cloud removal
• Semantic drift
• Downstream interpretation
• Remote sensing
• Interpretability

Innovation

• Observation-Anchored Residual Flow
• Geo-Contextual Prior Alignment
• Interpretation-Oriented Cloud Removal
• Vision Foundation Model
• Semantic Manifold