MoringaLeafNet: A multi-class leaf disease dataset for precision agriculture and deep learning research

Area of Science:

• Agricultural Science
• Plant Pathology
• Computer Vision

Background:

• Moringa Oleifera is globally valued for its nutritional and health benefits.
• Moringa cultivation faces challenges due to rapid spread of diseases like Yellow Leaf, Bacterial Leaf Spot, and Cercospora Leaf Spot.
• Disease outbreaks reduce crop yield, increase farmer costs through pesticide use, and harm the environment.

Purpose of the Study:

• To introduce the MoringaLeafNet dataset for early detection of Moringa leaf diseases.
• To facilitate the development of advanced agricultural disease diagnostic systems.
• To support real-time insights for farmers to improve decision-making and crop management.

Main Methods:

• Collected high-quality images of Moringa leaves across different disease classes (Healthy Leaf, Yellow Leaf, Bacterial Leaf Spot, Cercospora Leaf Spot).
• Acquired images from two nurseries in Bangladesh under diverse weather conditions.
• Applied data augmentation techniques including random rotation, flipping, and brightness/contrast adjustments for deep learning applications.

Main Results:

• The MoringaLeafNet dataset comprises images categorized into four distinct classes.
• Data augmentation ensures robustness and suitability for deep learning model training.
• The dataset serves as a foundational resource for developing automated Moringa disease identification tools.

Conclusions:

• The MoringaLeafNet dataset is crucial for advancing early recognition of Moringa leaf diseases.
• This resource can significantly contribute to the development of AI-powered agricultural diagnostic systems.
• Empowering farmers with timely disease detection can lead to improved crop yields and sustainable farming practices.