ExtremeEarth Polar Use Case Training Data
This is a set of sea ice, and iceberg, analysis data prepared for the ExtremeEarth project. The intention is that this dataset is useful for training and validating automated methods for processing satellite images, although it could also be applied to ice analyst training. The region chosen for the...
| Main Author: | |
|---|---|
| Format: | Dataset |
| Language: | English |
| Published: |
Zenodo
2020
|
| Subjects: | |
| Online Access: | https://dx.doi.org/10.5281/zenodo.3695276 https://zenodo.org/record/3695276 |
| Summary: | This is a set of sea ice, and iceberg, analysis data prepared for the ExtremeEarth project. The intention is that this dataset is useful for training and validating automated methods for processing satellite images, although it could also be applied to ice analyst training. The region chosen for the analysis was Danmarkshavn on the east coast of Greenland. This has a wide variety of sea ice and iceberg conditions, and was continuously monitored by key European Sentinel satellites during 2018. The dataset consists of 12 days, approximately monthly, throughout the year, covering winter and spring, summer melt, and autumn freeze-up. The presence of extensive (land) fast ice in the area ensures that classifications on those areas can be applied to additional dates. The sea ice classification includes primary and secondary ice types, including partial concentrations and form (floe size). The initial version contains basic ice/water classification, and otherwise the polygons remain unclassified for sea ice concentrations, ice type and form. This information will be added in later versions. |
|---|