Marine Underwater Environment Database (MUED)

In 2018 the Marine Underwater Environment Database (MUED) was developed as a benchmark repository for assisting research in vision saliency detection for underwater imagery and promoting research on underwater image processing and computer vision methodology. The developers used a cube pool to create a controlled sea-mimicking environment, equipped with a light controller and camera to create different oceanic conditions impacting image quality. With this they created the database, containing nearly 9000 underwater images of over 400  individual groups of conspicuous objects with complex backgrounds, multiple salient objects, and complicated variations in pose, spatial location, illumination, and water cloudiness (Jian et al., 2018, p. 1). The images were all manually labeled, with ground-truth information included. These images can now be used for reference in object identification, motion detection, masking, and classification methods by in-field researchers puzzling over their underwater imagery. 


Jian, M., Qi, Q., Yu, H., Dong, J., Cui, C., Nie, X., … Lam, K.-M. (2019). The extended marine underwater environment database and baseline evaluations. Applied Soft Computing, 80, 425–437.


See also:


National Oceanic and Atmospheric Administration. (2018, July 11). How much of the ocean have we explored? Retrieved May 20, 2020, from


Biodiversity, Climate Change, Data, Ecological Modelling, Ecological Monitoring, Visual Technologies