Tea-Leaf Disease and Pest Detection Dataset with 9,591 Annotated Images

9,591 annotated images of tea leaves across eight visual categories, including healthy leaves and seven disease or pest infestation types. The dataset was compiled by Xiang Lv and accompanies a research paper on real-time tea disease detection, supporting exact reproduction of the reported results. It includes pre-split training, validation, and test sets totaling approximately 366 MB.

Use Cases

• Training object detection models for tea leaf disease classification based on the eight annotated visual categories.
• Benchmarking lightweight neural network architectures for real-time agricultural pest detection.
• Reproducing and validating the results of the YOLOv8n-based methodology described in the accompanying paper.
• Developing automated plant health monitoring systems based on annotated bounding boxes of diseases and pests.

Strengths

• Contains 9,591 annotated images, providing a substantial base for model training.
• Includes 7,260 training, 784 validation, and 1,547 test images with corresponding label files, facilitating immediate use.
• Annotations follow the standard YOLO format and the bundle is drop-in compatible with the Ultralytics YOLO training pipeline.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• The original images were compiled from a single online community, which may introduce source bias.
• The dataset's license provenance is noted as unclear from the original distributor, requiring user caution for redistribution.

Provenance

Source

Images originally compiled from publicly accessible tea-disease pages on the Tencent Developer Community.

Collection Method

Processed through the Roboflow data-preparation pipeline.

Freshness

Last updated 2026-05-12 12:59:53; freshness should be verified.