underwater 360 benchmark dataset
收藏资源简介:
该数据集名为underwater 360 benchmark dataset,由西南大学和石溪大学创建,旨在解决水下图像的几何和颜色校正问题。数据集包含五个实验设置,每个设置约有60个视图,总计约300条数据。数据集通过实验室环境捕捉,包含水下图像及其对应的空气中的真实图像,用于模型评估。创建过程中考虑了水下环境的复杂性,如水箱变形、悬浮物体和光照变化。该数据集主要应用于水下海洋学、考古学、导航增强等领域,旨在提高水下图像的质量和应用效果。
The dataset named Underwater 360 Benchmark Dataset was developed by Southwest University and Stony Brook University, aiming to address the geometric and color correction challenges for underwater images. It comprises five experimental setups, each containing approximately 60 views, with a total of around 300 data samples. Captured in a laboratory setting, the dataset includes underwater images and their corresponding ground-truth aerial images for model evaluation. The complexities inherent in underwater environments, such as tank distortion, suspended objects and illumination variations, were taken into account during the dataset creation. This dataset is mainly applied in fields like underwater oceanography, archaeology, navigation augmentation and other related areas, with the objective of improving the quality and application effects of underwater images.




