Datasets:

Project-AgML
/

rural_LUCC_segmentation

dataset_info:
  features:
    - name: image
      dtype: image
    - name: mask
      dtype: image
  splits:
    - name: train
      num_bytes: 16263869
      num_examples: 2042
  download_size: 17997550
  dataset_size: 16263869
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: apache-2.0
task_categories:
  - image-segmentation
size_categories:
  - 1K<n<10K

Rural Lucc Segmentation

This dataset provides RGB satellite imagery for semantic segmentation of rural land use and land cover change in Jiangning District, Nanjing City, China. Captured in 2024 from a satellite platform, the images depict field environments typical of agricultural landscapes under real-world conditions. The real-world data offers a valuable resource for computer vision research focused on rural land monitoring and analysis. The dataset contains 2,042 images with pixel-level mask annotations.

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

Citation

@article{zhao2025land,
  title={Land-Unet: A deep learning network for precise segmentation and identification of non-structured land use types in rural areas for green urban space analysis},
  author={Zhao, Yan and Xie, Junru and Zhu, Huiru and Luo, Taige and Xiong, Yao and Fan, Chenyang and Xia, Haoxiang and Chen, Yuheng and Zhang, Fuquan},
  journal={Ecological Informatics},
  volume={87},
  pages={103078},
  year={2025},
  publisher={Elsevier}
}

https://www.kaggle.com/datasets/vvghigh/ruraluse

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