SARU: A Shadow-Aware and Removal Unified Framework for Remote Sensing Images with New Benchmarks

2026-04-28

AI Summary

This study addresses the challenge of shadows in remote sensing imagery, which severely degrade visual quality and impair downstream task performance. The authors propose SARU, a unified two-stage framework: the first stage employs a dual-branch network (DBCSF-Net) to fuse multi-color-space and semantic features for generating high-accuracy shadow masks; the second stage leverages these masks with a training-free physical algorithm (N²SGSR) to transfer illumination information from neighboring non-shadow regions for single-image shadow removal. Notably, this work is the first to jointly model shadow detection and removal within a unified framework, thereby avoiding error propagation and eliminating the need for paired training data. The authors also introduce two new benchmark datasets, RSISD and SiSRB. Experiments demonstrate that SARU achieves state-of-the-art performance on AISD, RSISD, and SiSRB, significantly enhancing shadow removal quality and robustness in downstream tasks.

Abstract

Shadows are a prevalent problem in remote sensing imagery (RSI), degrading visual quality and severely limiting the performance of downstream tasks like object detection and semantic segmentation. Most prior works treat shadow detection and removal as separate, cascaded tasks, which can lead to cumbersome process and error accumulation. Furthermore, many deep learning methods rely on paired shadow and non-shadow images for training, which are often unavailable in practice. To address these challenges, we propose Shadow-Aware and Removal Unified (SARU) Framework , a cohesive two-stage framework. First, its dual-branch detection module (DBCSF-Net) fuses multi-color space and semantic features to generate high-fidelity shadow masks, effectively distinguishing shadows from dark objects. Then, leveraging these masks, a novel, training-free physical algorithm (N$^2$SGSR) restores illumination by transferring properties from adjacent non-shadow regions within the single input image. To facilitate rigorous evaluation and foster future work, we also introduce two new benchmark datasets: the RSI Shadow Detection (RSISD) dataset and the Single-image Shadow Removal Benchmark (SiSRB). Extensive experiments demonstrate that SARU achieves state-of-the-art performance on both the public AISD dataset and our newly introduced benchmarks. By holistically integrating shadow detection and removal to mitigate error propagation and eliminating the dependency on paired training data, SARU establishes a robust, practical framework for real-world RSI analysis. The source code and datasets are publicly available at: https://github.com/AeroVILab-AHU/SARU-Framework.

Problem

• shadow detection
• shadow removal
• remote sensing images
• unpaired data
• error propagation

Innovation

• shadow detection
• shadow removal
• unified framework
• training-free algorithm
• remote sensing images