Datasets:

Project-AgML
/

tea_sickness_classification

dataset_info:
  features:
    - name: image
      dtype: image
    - name: label
      dtype:
        class_label:
          names:
            '0': Anthracnose
            '1': algal leaf
            '2': bird eye spot
            '3': brown blight
            '4': gray light
            '5': healthy
            '6': red leaf spot
            '7': white spot
  splits:
    - name: train
      num_bytes: 780911770
      num_examples: 885
  download_size: 780957799
  dataset_size: 780911770
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: cc-by-4.0
task_categories:
  - image-classification
size_categories:
  - n<1K

Tea Sickness Classification

This dataset contains real field images of tea leaves affected by various diseases, collected in tea gardens across Anhui Province, China. Images were captured using a handheld iPhone 14 Pro Max with RGB imaging during October 2023, providing practical examples for agricultural disease detection research. The dataset contains 885 images across 8 classes: Anthracnose, algal leaf, bird eye spot, brown blight, gray light, healthy, red leaf spot, white spot.
Images per class:

Anthracnose: 100
algal leaf: 113
bird eye spot: 100
brown blight: 113
gray light: 100
healthy: 74
red leaf spot: 143
white spot: 142

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

Citation

@article{yang2025lightweight,
  title={Lightweight wavelet-CNN tea leaf disease detection},
  author={Yang, Jing and Xu, GaoJian and Yang, MengDao and Lin, ZhengPei},
  journal={PLOS One},
  volume={20},
  pages={e0323322},
  year={2025},
  publisher={Public Library of Science}
}

The dataset itself can be cited as:

Gibson Kimutai. (2022). tea sickness dataset [Dataset]. Mendeley. https://doi.org/10.17632/J32XDT2FF5.2

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