EuroSAT Dataset

The EuroSAT dataset consists of satellite imagery for land use and land cover classification. It contains labeled images of 10 different land cover classes.

Please see our GFM-Bench for more information about how to use the dataset! 🙂

Metadata

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

S2_MEAN = [1354.40546513, 1118.24399958, 1042.92983953, 947.62620298, 1199.47283961, 1999.79090914, 2369.22292565, 2296.82608323, 732.08340178, 12.11327804, 1819.01027855, 1118.92391149, 2594.14080798]
S2_STD = [245.71762908, 333.00778264, 395.09249139, 593.75055589, 566.4170017, 861.18399006, 1086.63139075, 1117.98170791, 404.91978886, 4.77584468, 1002.58768311, 761.30323499, 1231.58581042]

metadata = {
    "s2c": {
        "bands": ["B1", "B2", "B3", "B4", "B5", "B6", "B7", "B8", "B8A", "B9", "B10", "B11", "B12"],
        "channel_wv": [442.7, 492.4, 559.8, 664.6, 704.1, 740.5, 782.8, 832.8, 864.7, 945.1, 1373.5, 1613.7, 2202.4],
        "mean": S2_MEAN,
        "std": S2_STD,
    },
    "s1": {
        "bands": None,
        "channel_wv": None,
        "mean": None,
        "std": None   
    }
}

SIZE = HEIGHT = WIDTH = 64

NUM_CLASSES = 10

spatial_resolution = 10

Split

The EuroSAT dataset consists splits of:

• train: 16200 samples
• val: 5400 samples
• test: 5400 samples

Features:

The EuroSAT dataset consists of following features:

• optical: the Sentinel-2 image.
• label: the classification label.
• optical_channel_wv: the wavelength of each optical channel.
• spatial_resolution: the spatial resolution of images.

Citation

If you use the EuroSAT dataset in your work, please cite the original paper:

@article{helber2019eurosat,
  title={Eurosat: A novel dataset and deep learning benchmark for land use and land cover classification},
  author={Helber, Patrick and Bischke, Benjamin and Dengel, Andreas and Borth, Damian},
  journal={IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing},
  volume={12},
  number={7},
  pages={2217--2226},
  year={2019},
  publisher={IEEE}
}