WeatherBench
收藏资源简介:
WeatherBench数据集由大连理工大学的研究团队创建,旨在为所有类型的恶劣天气图像恢复提供真实世界的基准数据集。该数据集包含在各种天气条件下(如雨、雪、雾)以及多样化的室外场景和照明设置下捕获的图像对。数据集提供了精确对齐的退化图像和干净图像,可用于监督学习和严格的评估。数据集规模庞大,包含42,002对图像,涵盖了多种天气类型,并提供了白天和晚上的场景,从而提高了照明多样性。数据集的创建过程严格控制了外部环境因素和内部技术参数,以确保图像质量和对齐精度。WeatherBench数据集为推进鲁棒且实用的所有类型恶劣天气图像恢复在现实场景中的应用奠定了宝贵的基础。
The WeatherBench dataset was developed by a research team from Dalian University of Technology, with the goal of providing a real-world benchmark dataset for image restoration tasks targeting all types of severe weather. This dataset comprises image pairs captured under diverse weather conditions (e.g., rain, snow, fog), varied outdoor scenarios, and different lighting configurations. It offers precisely aligned degraded and clean images, which are suitable for supervised learning and rigorous model evaluation. Boasting a large scale, the dataset contains 42,002 image pairs, covering multiple weather types and encompassing both daytime and nighttime scenes, thus enhancing the diversity of lighting conditions. The creation process of the dataset strictly regulates external environmental factors and internal technical parameters to guarantee the image quality and alignment precision. The WeatherBench dataset lays a valuable foundation for promoting the real-world deployment of robust and practical image restoration methods for all kinds of severe weather.




