---
dataset_info:
  features:
  - name: rgb
    dtype: image
  - name: mask
    dtype: image
  - name: split
    dtype: string
  splits:
  - name: train
    num_bytes: 3092369713
    num_examples: 213758
  download_size: 3096602246
  dataset_size: 3092369713
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
license: cc-by-4.0
task_categories:
- image-segmentation
size_categories:
- 100K<n<1M
---
# Sen2 Lulc

The Sen2_LULC dataset provides real multispectral satellite imagery for semantic segmentation of land use and land cover in the central Indian region. Captured by Sentinel-2 at 10-meter resolution during February-March 2021, the data was collected from a satellite platform over field environments, offering high-quality inputs for agricultural and environmental monitoring applications. The dataset contains 213,758 images with pixel-level mask annotations.

This dataset is indexed on https://project-agml.github.io/ as part of the AgML python library.

The original train/test/val split has been preserved in the `split` column.

## Citation

```bibtex
@article{sawant2023sen,
  title={Sen-2 LULC: Land use land cover dataset for deep learning approaches},
  author={Sawant, Suraj and Garg, Rahul Dev and Meshram, Vishal and Mistry, Shrayank},
  journal={Data in Brief},
  volume={51},
  pages={109724},
  year={2023},
  publisher={Elsevier}
}
```

Sawant, Suraj; Garg, Rahul Dev; Meshram, Vishal; Mistry, Shrayank (2023), “Sen-2 LULC ”, Mendeley Data, V3, doi: 10.17632/f4ky6ks248.3

*This dataset was reformatted from its original format to match HuggingFace standards.*