vapaau/THEBUCKETv1

The BUCKETv1 Datset

This dataset is published as part of our paper:
Sea-ing Through Scattered Rays: Revisiting the Image Formation Model for Realistic Underwater Image Generation

Accepted in the CVAUI & AAMVEMV workshop in ICCV 2025

• Project page: https://vap.aau.dk/sea-ing-through-scattered-rays/
• Code: https://github.com/vismiroglou/STSR
• arXiv: https://arxiv.org/abs/2509.15011

Dataset Description:

The purpose of this dataset is to study the effects of turbidity on computer vision models by inducing turbidity in a controlled environment.

The setup involved four GoPro cameras, equally spaced along the perimeter of a cylindrical tank with opaque walls. Data collection took place in a dark room, and lighting was introduced through four Blue Robotics Lumen subsea lights and a diffused lamp centered above the tank. Turbidity was induced using increasing amounts of oat milk and clay, and was accurately measured with a turbidimeter.

Data collection setup	Camera-light setup

On the bottom of the tank, sets of objects were placed on a LEGO baseplate, retaining object positions. This ensured the preservation of the scene between the clear reference images and the ones collected as turbidity increases. Objects include, but are not limited to, artificial and real rocks, as well as trash of different shapes and colors.

Example clear image	Example turbid image

64 images of varying turbidity with ambient light were used for paper experiments. This version of the dataset includes an additional 36 for a total of 90 images. The measured turbidity in SNU is provided in .csv files for each instance. There are multiple measurements of different samples at each turbidity level.

Citation:

@misc{ismiroglou2025seaingscatteredraysrevisiting,
      title={Sea-ing Through Scattered Rays: Revisiting the Image Formation Model for Realistic Underwater Image Generation}, 
      author={Vasiliki Ismiroglou and Malte Pedersen and Stefan H. Bengtson and Andreas Aakerberg and Thomas B. Moeslund},
      year={2025},
      eprint={2509.15011},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2509.15011}, 
}