Published July 22, 2018 | Version v2
Dataset Open

EuroSAT: A Novel Dataset and Deep Learning Benchmark for Land Use and Land Cover Classification

Authors/Creators

Description

EuroSAT is a land use and land cover classification dataset. The dataset is based on Sentinel-2 satellite imagery covering 13 spectral bands and consists of 10 LULC classes with a total of 27,000 labeled and geo-referenced images. The dataset is associated with the publications "Introducing EuroSAT: A Novel Dataset and Deep Learning Benchmark for Land Use and Land Cover Classification" and "EuroSAT: A Novel Dataset and Deep Learning Benchmark for Land Use and Land Cover Classification".

EuroSAT_RGB.zip contains the RGB version of the dataset, which includes the optical R, G and B frequency bands encoded as JPEG images.

EuroSAT_MS.zip contains the multi-spectral version of the EuroSAT dataset, which includes all 13 Sentinel-2 bands in the original value range.

Notes

Sentinel data is free and open to the public under EU law. Please consider the Copernicus Sentinel Data Terms and Conditions (https://sentinel.esa.int/documents/247904/690755/Sentinel_Data_Legal_Notice) when using Copernicus Sentinel data. The work associated with the creation of the dataset and the publication of the related article and conference paper was partially funded by the BMBF project DeFuseNN (01IW17002). The authors thank NVIDIA for the support within the NVIDIA AI Lab program.

Files

EuroSAT_MS.zip

Files (2.2 GB)

Name Size
md5:091174add3c8e680a49244acf185b9f0
2.1 GB Preview Download
md5:f46e308c4d50d4bf32fedad2d3d62f3b
94.7 MB Preview Download

Additional details

Related works

Is described by
Journal article: 10.1109/JSTARS.2019.2918242 (DOI)
Conference paper: 10.1109/IGARSS.2018.8519248 (DOI)
Is referenced by
https://github.com/phelber/EuroSAT (URL)