UAV-Rain1k: A Benchmark for Raindrop Removal from UAV Aerial Imagery

TL;DR

This work tackles raindrop removal in UAV aerial imagery by introducing UAV-Rain1k, a UAV-specific raindrop benchmark built from Blender-based raindrop generation and diverse, high-quality background collection. The authors propose a realistic synthesis pipeline where raindrop effects combine with clean backgrounds via $R_d = (1-M) \odot B + D$, and they perform subjective assessments to validate realism. Six state-of-the-art deraining methods are benchmarked, with Transformer-based models achieving top PSNR/SSIM at the cost of higher computational complexity, and downstream object detection (YOLOv8) showing varied gains depending on the method. Overall, UAV-Rain1k provides a realistic, scalable testbed that reveals limitations of current approaches and guides future work toward efficient, joint low- and high-level vision tasks in UAV contexts.

Abstract

Raindrops adhering to the lens of UAVs can obstruct visibility of the background scene and degrade image quality. Despite recent progress in image deraining methods and datasets, there is a lack of focus on raindrop removal from UAV aerial imagery due to the unique challenges posed by varying angles and rapid movement during drone flight. To fill the gap in this research, we first construct a new benchmark dataset for removing raindrops from UAV images, called UAV-Rain1k. In this letter, we provide a dataset generation pipeline, which includes modeling raindrop shapes using Blender, collecting background images from various UAV angles, random sampling of rain masks and etc. Based on the proposed benchmark, we further present a comprehensive evaluation of existing representative image deraining algorithms, and reveal future research opportunities worth exploring. The proposed dataset is publicly available at https://github.com/cschenxiang/UAV-Rain1k.

Figures (8)

• Figure 1: Classification diagram of publicly available datasets for haze removal and rain removal. Our proposed UAV-Rain1k fills the research gap.
• Figure 2: Illustration of the dataset generation pipeline.
• Figure 3: Distribution of rain and scene of the proposed benchmark.
• Figure 4: User study results. The ratings given by all participants on different raindrop datasets.
• Figure 5: Example images from previous representative datasets (Raindrop qian2018attentive and RainDS quan2021removing) and our proposed more challenging UAV-Rain1k.