The TURBID, is an open image dataset that has been generated to contribute with the underwater research area. TURBID consists in a collection of five different subsets of degraded images with its respective ground-truth.

The subset Milk and DeepBlue has 20 images each and the subset Chlorophyll has 42 images.

The subsets Blue and Green Tea will be available soon...

Subset Milk

Subset Chlorophyll

Subset Clhlorophyll

Subset DeepBlue

Subset Milk

Subset Chlorophyll

Subset Chlorophyll

Subset DeepBlue

More images will be available soon...


Want to use the TURBID dataset? You can download it from here:

Complete Dataset



If you use the any image from the TURBID dataset, please cite the following work:

A Dataset to Evaluate Underwater Image Restoration Methods. IEEE OCEANS 2016-Shanghai, 2016.



title={A dataset to evaluate underwater image restoration methods},

author={Duarte, Amanda and Codevilla, Felipe and Gaya, Joel De O and Botelho, Silvia SC},

booktitle={OCEANS 2016-Shanghai},




About Us

We are the research team who developed the TURBID dataset. We are part of the Robotics and Intelligent Automation Group - NAUTEC at the Federal University of Rio Grande - FURG (Universidade Federal do Rio Grande - FURG) in Brazil. Our research group has been working on different aspects of computer vision, underwater robotics, and machine learning. With this dataset, we aim to collaborate with the research area focused on underwater environments bringing new underwater data with its respective ground-truth.