adamkatchee/grape-leaf-disease-augmented-dataset

🍇🍃 Grape Leaf Disease Augmented Dataset`

Information

This dataset consists of 18,054 grape leaf images across 4 class (ESCA, Leaf Blight, Black Rot, Healthy) for use in image classification tasks.

Dataset has been created as a modification to the original dataset from the PlantVillage dataset: https://www.kaggle.com/datasets/abdallahalidev/plantvillage-dataset

The images have been used from this dataset: https://www.kaggle.com/datasets/rm1000/grape-disease-dataset-original/data

This dataset has been created and used as a part of the work on a bachelor thesis.

---

Supported Tasks

• image-classification

---

Dataset Structure

The dataset consists of 4 classes, with the following distribution across the training and test sets:

Class Distribution

Class Name	Train	Test	Total
Black Rot	3,776	944	4,720
ESCA	3,840	960	4,800
Leaf Blight	3,444	860	4,304
Healthy	3,384	846	4,230
Total	14,444	3,610	18,054

Dataset Creation

Original Sources

• https://www.kaggle.com/datasets/abdallahalidev/plantvillage-dataset
• https://www.kaggle.com/datasets/rm1000/grape-disease-dataset-original/data

Creation Process

This dataset has been created using OpenCV Python library. Background pictures were taken in the region of Bratislava, Slovakia in January 2025.