metadata

license: cc-by-nc-2.0
task_categories:
  - object-detection
size_categories:
  - 1K<n<10K

Locating and measuring marine aquaculture production with remote sensing: a computer vision approach in the French Mediterranean

Sebastian Quaade, Andrea Vallebueno, Olivia D. N. Alcabes, Kit T. Rodolfa, Daniel E. Ho

This is the data repository for the manuscript "Locating and measuring marine aquaculture production with remote sensing: a computer vision approach in the French Mediterranean."

Overview

To enable other researchers seeking to apply our method, we make available an object detection dataset (in YOLO format) comprising human-annotated finfish cage bounding boxes. The dataset comprises 286 images sourced from aerial imagery of the French Mediterranean that was derived from Institut national de l'information géographique et forestière's (IGN) BD Ortho series. The dataset contains 4,142 finfish cage bounding boxes. The dataset comprises two classes: circular cages (567 labels) and square cages (3,575 labels).

Data description

images/ Includes the 286 images as .jpeg files
labels/ Includes the labels for each image in YOLO format, including the following classes: circular cages (class 0) and square cages (class 1)

Citation (TBD)

@article{citation_key,
  author = {Author Name},
  title = {Article Title},
  journal = {Journal Name},
  year = {Year},
  volume = {Volume},
  number = {Number},
  pages = {Pages},
  doi = {DOI}
}

References

Institut national de l’information géographique et forestière, BD704 ORTHO, version 2.0, République Française Géoservices (2023); https://geoservices.ign.fr/documentation/donnees/ortho/bdortho.