Seaview Survey Photo-quadrat and Image Classification Dataset

The primary scientific dataset arising from the XL Catlin Seaview Survey project is the “Seaview Survey Photo-quadrat and Image Classification Dataset”, consisting of: (1) over one million standardised, downward-facing “photo-quadrat” images covering approximately 1m2 of the sea floor; (2) human-classified annotations that can be used to train and validate image classifiers; (3) benthic cover data arising from the application of machine learning classifiers to the photo-quadrats; and (4) the triplets of raw images (covering 360o) from which the photo-quadrats were derived.Photo-quadrats were collected between 2012 and 2018 at 860 transect locations around the world, including: the Caribbean and Bermuda, the Indian Ocean (Maldives, Chagos Archipelago), the Coral Triangle (Indonesia, Philippines, Timor-Leste, Solomon Islands), the Great Barrier Reef, Taiwan and Hawaii.For additional information regarding methodology, data structure, organization and size, please see attached document “Dataset documentation”.