Underwater Visual and Visual-Inertial Datasets (UVVID)

dataset

posted on 2024-12-06, 07:55 authored by Fletcher ThompsonFletcher Thompson, David Dylan O’Brien-MøllerDavid Dylan O’Brien-Møller, Patrizio MarianiPatrizio Mariani, Bo LundgrenBo Lundgren

UVVID

The Underwater Visual and Visual-Inertial Dataset (UVVID) contains various forms of underwater video imagery of roughly planar environments (walls and seabottom). Datasets are split between visual-only (i.e. just video data), or visual-inertial (i.e. including IMU stabilization files for the cameras). The datasets are provided as-is for testing and development of algorithms using concepts in visual odometry, SLAM, mosaicking and other mapping applications.

Datasets also include post-processed forms of external data sources, such as AQUALOC and the Underwater Caves Dataset. These are provided as updated versions of the datasets that include additional transformations and sensor fusion methods to reduce ambiguity in the preprocessing methods supporting the target applications. Please see the README.md files in each subfolder for descriptions of specific datasets.

References

Ferrera M, Creuze V, Moras J, Trouvé-Peloux P. AQUALOC: An underwater dataset for visual–inertial–pressure localization. The International Journal of Robotics Research. 2019;38(14):1549-1559. doi:10.1177/0278364919883346

Mallios, A.; Vidal, E.; Campos, R. & Carreras, M. Underwater caves sonar data set. The International Journal of Robotics Research, 2017, 36, 1247-1251
doi: 10.1177/0278364917732838