Datasets:

Project-AgML
/

simta_rice_mapping_classification

dataset_info:
  features:
    - name: bands
      dtype: binary
    - name: bands_shape
      list: int64
    - name: bands_dtype
      dtype: string
    - name: band_metadata
      dtype: string
    - name: band_order
      list: string
    - name: label
      dtype:
        class_label:
          names:
            '0': 2019Canifornia
            '1': EImage
            '2': ThImage
            '3': TImage
  splits:
    - name: train
      num_bytes: 5400475005
      num_examples: 3024
  download_size: 4236996494
  dataset_size: 5400475005
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: cc-by-4.0
task_categories:
  - image-classification
size_categories:
  - 1K<n<10K

Simta Rice Mapping Classification

This dataset provides real SAR satellite imagery of rice fields collected across the Arkansas River Basin, Sacramento Valley, and Suihua regions using Sentinel-1 between 2017 and 2019. It captures multi-temporal rice cultivation patterns in agricultural field environments across the United States and China, offering valuable data for computer vision research in crop monitoring and mapping. The dataset contains 3,024 images across 4 classes: 2019Canifornia, EImage, ThImage, TImage.
Images per class:

2019Canifornia: 375
EImage: 706
ThImage: 884
TImage: 1,059

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

Citation

@article{ren2026simta,
  title={SimTA: A Dual-Polarization SAR Time-Series Rice Field Mapping Model Based on Deep Feature-Level Fusion and Spatiotemporal Attention},
  author={Ren, Dong and Liang, Jiaxuan and Liu, Li and Wei, Pengliang and Yang, Lingbo and Wang, Lu and Sun, Hang and Zhang, Kehan and Qiu, Bingwen and Liu, Weiwei and Huang, Jingfeng},
  journal={Remote Sensing},
  volume={18},
  pages={1237},
  year={2026},
  publisher={MDPI}
}

The dataset itself can be cited as:

Liu, L., Deng, X., & Quan, T. (2025). Rice Mapping Based on SAR Imagery – Arkansas River Basin, Sacramento Region [Dataset]. Zenodo. https://doi.org/10.5281/ZENODO.17997718

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