BigEarthNet

BigEarthNet is a large-scale benchmark dataset for multi-label classification, derived from Sentinel-1 (radar) and Sentinel-2 (optical) satellite imagery.

We have pre-processed the dataset by upsampling all sentinel-2 channels to 120x120 pixels and concatenated them together. Please see Torchgeo/bigearthnet for more information about pre-processing. In addition, we map the original 43 land cover classes to 19 broader categories using a predefined conversion scheme.

How to Use This Dataset

from datasets import load_dataset

dataset = load_dataset("GFM-Bench/BigEarthNet")

Also, please see our GFM-Bench repository for more information about how to use the dataset! 🤗

Dataset Metadata

The following metadata provides details about the Sentinel-2 imagery used in the dataset:

• Number of Sentinel-1 Bands: 2
• Sentinel-1 Bands: VV, VH
• Number of Sentinel-2 Bands: 12
• Sentinel-2 Bands: B01 (Coastal aerosol), B02 (Blue), B03 (Green), B04 (Red), B05 (Vegetation red edge), B06 (Vegetation red edge), B07 (Vegetation red edge), B08 (NIR), B8A (Narrow NIR), B09 (Water vapour), B11 (SWIR), B12 (SWIR)
• Image Resolution: 120 x 120 pixels
• Spatial Resolution: 10 meters
• Number of Classes: 19
• Class Labels:
  • Urban fabric
  • Industrial or commercial units
  • Arable land
  • Permanent crops
  • Pastures
  • Complex cultivation patterns
  • Land principally occupied by agriculture, with significant areas of natural vegetation
  • Agro-forestry areas
  • Broad-leaved forest
  • Coniferous forest
  • Mixed forest
  • Natural grassland and sparsely vegetated areas
  • Moors, heathland and sclerophyllous vegetation
  • Transitional woodland, shrub
  • Beaches, dunes, sands
  • Inland wetlands
  • Coastal wetlands
  • Inland waters
  • Marine waters

Dataset Splits

The BigEarthNet dataset consists following splits:

• train: 269,695 samples
• val: 123,723 samples
• test: 125,866 samples

Dataset Features:

The BigEarthNet dataset consists of following features:

• radar: the Sentinel-1 image.
• optical: the Sentinel-2 image.
• label: the classification label.
• radar_channel_wv: the central wavelength of each Sentinel-1 bands.
• optical_channel_wv: the central wavelength of each Sentinel-2 bands.
• spatial_resolution: the spatial resolution of images.

Citation

If you use the BigEarthNet dataset in your work, please cite original papers:

@inproceedings{sumbul2019bigearthnet,
  title={Bigearthnet: A large-scale benchmark archive for remote sensing image understanding},
  author={Sumbul, Gencer and Charfuelan, Marcela and Demir, Beg{\"u}m and Markl, Volker},
  booktitle={IGARSS 2019-2019 IEEE International Geoscience and Remote Sensing Symposium},
  pages={5901--5904},
  year={2019},
  organization={IEEE}
}