BigEarthNet Benchmark Archive Now Available on Radiant MLHub, the Open Repository for Geospatial Training Data
BigEarthNet is a new large-scale benchmark geospatial training data consisting of multi-label land cover classes in ten European countries.
Radiant Earth Foundation, the leader in enabling access to geospatial training data, is pleased to announce the availability of the BigEarthNet large-scale benchmark archive through Radiant MLHub, the world’s first open library dedicated to Earth observation (EO) training data.
The BigEarthNet archive consists of 590, 326 Sentinel-2 image patches with spectral bands at 10, 20, and 60-meter resolution. The satellite images were acquired in different seasons between June 2017 and May 2018 over Austria, Belgium, Finland, Ireland, Kosovo, Lithuania, Luxembourg, Portugal, Serbia, and Switzerland. Each patch is annotated with multiple land cover labels documenting the spatial distribution of every land cover class across the dataset region. The majority of image patches contain at most five labels, including continuous urban fabric, green urban areas, mixed forests, discontinuous urban fabric, non-irrigated arable land, among 43 other classes from the 2018 CORINE Land Cover (CLC) inventory.
The BigEarthNet archive is significantly larger than any existing public training dataset and opens up promising opportunities to advance research for the analysis of large-scale geospatial image archives. The Remote Sensing Image Analysis (RSiM) Group and the Database Systems and Information Management (DIMA) Group at the Technische Universität Berlin (TU Berlin) constructed the BigEarthNet archive. This work is supported by the European Research Council under the ERC Starting Grant BigEarth and by the German Ministry for Education and Research as Berlin Big Data Center (BBDC).
The addition of BigEarthNet to Radiant MLHub data supports the community’s vision of creating a global repository of geospatial training data that can be leveraged to solve real world problems. Radiant MLHub is a cloud-based interoperable solution for registering and accessing geospatial training data. Shared data are accessible via a standardized API, and can, therefore, move across organizations, governments, and sectors to unlock new opportunities for data-based insights. By cataloging the BigEarthNet dataset in the SpatioTemporal Asset Catalog (STAC) specification, Radiant MLHub enables the search and filtering of the training data based on geographic region, land cover labels, cloud and cloud shadow presence, as well as a variety of other properties.