bigearthnet

  • Description:

The BigEarthNet is a new large-scale Sentinel-2 benchmark archive, consisting of 590,326 Sentinel-2 image patches. The image patch size on the ground is 1.2 x 1.2 km with variable image size depending on the channel resolution. This is a multi-label dataset with 43 imbalanced labels.

To construct the BigEarthNet, 125 Sentinel-2 tiles acquired between June 2017 and May 2018 over the 10 countries (Austria, Belgium, Finland, Ireland, Kosovo, Lithuania, Luxembourg, Portugal, Serbia, Switzerland) of Europe were initially selected. All the tiles were atmospherically corrected by the Sentinel-2 Level 2A product generation and formatting tool (sen2cor). Then, they were divided into 590,326 non-overlapping image patches. Each image patch was annotated by the multiple land-cover classes (i.e., multi-labels) that were provided from the CORINE Land Cover database of the year 2018 (CLC 2018).

Bands and pixel resolution in meters:

  • B01: Coastal aerosol; 60m
  • B02: Blue; 10m
  • B03: Green; 10m
  • B04: Red; 10m
  • B05: Vegetation red edge; 20m
  • B06: Vegetation red edge; 20m
  • B07: Vegetation red edge; 20m
  • B08: NIR; 10m
  • B09: Water vapor; 60m
  • B11: SWIR; 20m
  • B12: SWIR; 20m
  • B8A: Narrow NIR; 20m

License: Community Data License Agreement - Permissive, Version 1.0.

URL: http://bigearth.net/

Split Examples
'train' 590,326
  • Citation:
@article{Sumbul2019BigEarthNetAL,
  title={BigEarthNet: A Large-Scale Benchmark Archive For Remote Sensing Image Understanding},
  author={Gencer Sumbul and Marcela Charfuelan and Beg{"u}m Demir and Volker Markl},
  journal={CoRR},
  year={2019},
  volume={abs/1902.06148}
}

bigearthnet/rgb (default config)

  • Config description: Sentinel-2 RGB channels

  • Dataset size: 14.07 GiB

  • Feature structure:

FeaturesDict({
    'filename': Text(shape=(), dtype=string),
    'image': Image(shape=(120, 120, 3), dtype=uint8),
    'labels': Sequence(ClassLabel(shape=(), dtype=int64, num_classes=43)),
    'metadata': FeaturesDict({
        'acquisition_date': Text(shape=(), dtype=string),
        'coordinates': FeaturesDict({
            'lrx': int64,
            'lry': int64,
            'ulx': int64,
            'uly': int64,
        }),
        'projection': Text(shape=(), dtype=string),
        'tile_source': Text(shape=(), dtype=string),
    }),
})
  • Feature documentation:
Feature Class Shape Dtype Description
FeaturesDict
filename Text string
image Image (120, 120, 3) uint8
labels Sequence(ClassLabel) (None,) int64
metadata FeaturesDict
metadata/acquisition_date Text string
metadata/coordinates FeaturesDict
metadata/coordinates/lrx Tensor int64
metadata/coordinates/lry Tensor int64
metadata/coordinates/ulx Tensor int64
metadata/coordinates/uly Tensor int64
metadata/projection Text string
metadata/tile_source Text string

Visualization

bigearthnet/all

  • Config description: 13 Sentinel-2 channels

  • Dataset size: 176.63 GiB

  • Feature structure:

FeaturesDict({
    'B01': Tensor(shape=(20, 20), dtype=float32),
    'B02': Tensor(shape=(120, 120), dtype=float32),
    'B03': Tensor(shape=(120, 120), dtype=float32),
    'B04': Tensor(shape=(120, 120), dtype=float32),
    'B05': Tensor(shape=(60, 60), dtype=float32),
    'B06': Tensor(shape=(60, 60), dtype=float32),
    'B07': Tensor(shape=(60, 60), dtype=float32),
    'B08': Tensor(shape=(120, 120), dtype=float32),
    'B09': Tensor(shape=(20, 20), dtype=float32),
    'B11': Tensor(shape=(60, 60), dtype=float32),
    'B12': Tensor(shape=(60, 60), dtype=float32),
    'B8A': Tensor(shape=(60, 60), dtype=float32),
    'filename': Text(shape=(), dtype=string),
    'labels': Sequence(ClassLabel(shape=(), dtype=int64, num_classes=43)),
    'metadata': FeaturesDict({
        'acquisition_date': Text(shape=(), dtype=string),
        'coordinates': FeaturesDict({
            'lrx': int64,
            'lry': int64,
            'ulx': int64,
            'uly': int64,
        }),
        'projection': Text(shape=(), dtype=string),
        'tile_source': Text(shape=(), dtype=string),
    }),
})
  • Feature documentation:
Feature Class Shape Dtype Description
FeaturesDict
B01 Tensor (20, 20) float32
B02 Tensor (120, 120) float32
B03 Tensor (120, 120) float32
B04 Tensor (120, 120) float32
B05 Tensor (60, 60) float32
B06 Tensor (60, 60) float32
B07 Tensor (60, 60) float32
B08 Tensor (120, 120) float32
B09 Tensor (20, 20) float32
B11 Tensor (60, 60) float32
B12 Tensor (60, 60) float32
B8A Tensor (60, 60) float32
filename Text string
labels Sequence(ClassLabel) (None,) int64
metadata FeaturesDict
metadata/acquisition_date Text string
metadata/coordinates FeaturesDict
metadata/coordinates/lrx Tensor int64
metadata/coordinates/lry Tensor int64
metadata/coordinates/ulx Tensor int64
metadata/coordinates/uly Tensor int64
metadata/projection Text string
metadata/tile_source Text string