Melt pond from aerial photographs of the Healy–Oden Trans Arctic Expedition (HOTRAX)

• Main Authors: Ivan Sudakow, Vijayan Asari, Ruixu Liu, Denis Demchev
• Format: Dataset
• Language: English
• Published: 2022
• Subjects: melt pond; detection; HOTRAX; climate; machine learning; Arctic; Sea ice
• Online Access: https://zenodo.org/record/6602409; https://doi.org/10.5281/zenodo.6602409

The dataset contains mapped melt pond zones from aerial photographs which were obtained during helicopter photography flights as part of the Healy–Oden Trans Arctic Expedition (HOTRAX) between 5 August and 30 September 2005. The detection has been done using an original detection algorithm based on machine learning (U-Net). The areas in the images were divided into four classes: sea ice/snow, melt ponds, sub-merged ice and open water. The filenames contain IDs from the original dataset [Perovich et al 2008]. This work was supported by the Division of Physics at the U.S. National Science Foundation (NSF) through Grant PHY-2102906