-> NEREON dataset <-

In the context of the ATLANTIS project, the NEREON dataset (uNderwater datasEt for monoculaR dEpth estimatiON) was created to provide information for training Deep Learning Networks which aimed at estimating depth from single images in underwater environments. The dataset was acquired at the ATLANTIS Test Centre in Viana do Castelo, around the DURIUS buoy, designed for testing robotic solutions for O&M. The collection process resulted in close to 7000 images and point clouds (which were projected to match the resolution of the images), that depict the DURIUS structure as well as the its surrounding underwater environment.

The present repository contains the following items:

-> images.zip - Folder containing the dataset images (.png)
-> point_clouds.zip - Folder containing the point cloud files (.tiff)
-> intrinsic_params.zip - File containing the parameters for camera intrinsic calibration and 3D reconstruction of the projected point clouds
-> sample.zip - Folder containing some images and point cloud samples from the dataset

Authors: João Dionísio, Pedro Almeida, Pedro Nuno Leite, Andry Maykol Pinto
Contact information: joao.m.dionisio@inesctec.pt