---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: label
    dtype:
      class_label:
        names:
          '0': Forest
          '1': River
          '2': Highway
          '3': AnnualCrop
          '4': SeaLake
          '5': HerbaceousVegetation
          '6': Industrial
          '7': Residential
          '8': PermanentCrop
          '9': Pasture
  splits:
  - name: train
    num_bytes: 73654547.8
    num_examples: 21600
  - name: validation
    num_bytes: 9213645.6
    num_examples: 2700
  - name: test
    num_bytes: 9201991.7
    num_examples: 2700
  download_size: 91902630
  dataset_size: 92070185.1
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
  - split: test
    path: data/test-*
license: mit
size_categories:
- 10K<n<100K
task_categories:
- image-classification
---
# EuroSAT-RGB Dataset

### Dataset Description

The dataset comprises JPEG composite chips extracted from Sentinel-2 satellite imagery,
representing the Red, Green, and Blue bands. 
It encompasses 27,000 labeled and geo-referenced images across 10 Land Use and Land Cover (LULC) classes

## Dataset Structure

Splits : Train 80% Validation 10% Test 10% 
(Kept the original dataset's label distribution consistent in each split)

## Citation 

Helber, P., Bischke, B., Dengel, A., & Borth, D. (2018). EuroSAT: A Novel Dataset and Deep Learning Benchmark for Land Use and Land Cover Classification [Data set]. In EuroSAT: A Novel Dataset and Deep Learning Benchmark for Land Use and Land Cover Classification (Vol. 12, Number 7, pp. 2217–2226). Zenodo. https://doi.org/10.5281/zenodo.7711810