Interpretation-Oriented Cloud Removal via Observation-Anchored Residual Flow with Geo-Contextual Alignment
2026-07-02 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors developed a new method called Geo-Anchored Cloud Removal (GACR) that removes clouds in satellite images while keeping important details needed for further analysis. Their approach uses a special flow technique, Observation-Anchored Residual Flow (OAR-Flow), which grounds the cloud removal process more realistically by starting from the cloudy image rather than random noise. They also use a Geo-Contextual Prior Alignment (GCPA) to ensure the results stay true to real landscape features important for tasks like detecting land changes or segmenting images. Tests show their method improves both image quality and the accuracy of these follow-up tasks.
cloud removalremote sensingresidual flowsemantic segmentationchange detectionvision foundation modelgeospatial analysisimage reconstructionsemantic driftdownstream tasks
Authors
Ziyao Wang, Maonan Wang, Yucheng He, Xianping Ma, Ziyi Wang, Hongyang Zhang, Yirong Cheng, Man-on Pun
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.