MangoLeaf

Data Description
This dataset provides a comprehensive collection of mango leaf images classified into four categories: Healthy, Anthracnose, Gall Midge, and Powdery Mildew. It is specifically designed to facilitate research in plant disease recognition, deep learning-based image classification, and precision agriculture.

A total of 2,000 original images were acquired, with 500 images per class. To increase diversity and support the training of high-performing machine learning models, data augmentation techniques—including rotation, flipping, zooming, brightness adjustment, and contrast variation—were applied. This yielded an additional 1,500 augmented images per class, resulting in a complete dataset of 8,000 images (2,000 original + 6,000 augmented).

Class Definitions:
Healthy: Mango leaves exhibiting no visible symptoms, representing the non-infected control group.

Anthracnose: A common fungal disease caused by Colletotrichum species, typically manifested through dark, sunken lesions and necrotic spots.

Gall Midge: Caused by Procontarinia matteiana, resulting in gall formation, leaf distortion, and blister-like symptoms on the leaf surface.

Powdery Mildew: A fungal infection due to Oidium mangiferae, characterized by white, powdery growth, often leading to curling and premature leaf aging.

Image Acquisition:
Images were collected from multiple mango orchards using both DSLR and smartphone cameras under natural lighting and varied environmental conditions. The diversity in background, lighting, and leaf orientation enhances the robustness of the dataset for real-world applications.

Potential Applications:
This dataset serves as a valuable resource for:

Developing and evaluating deep learning models for plant disease detection

Building mobile applications for on-field mango disease diagnosis

Advancing smart farming and AI-driven agricultural solutions

Supporting computer vision research in biological and environmental domains

All images are manually annotated and reviewed by agricultural experts to ensure high-quality labeling. The dataset is released for non-commercial, academic, and research purposes to promote open science in the agricultural AI community.