Banana Leaves Imagery Dataset.

Area of Science:

• Agricultural Science
• Computer Science
• Plant Pathology

Background:

• Banana cultivation is vital globally, but susceptible to diseases like Black Sigatoka and Fusarium Wilt Race 1.
• Accurate and early disease diagnosis is crucial for effective crop management and yield preservation.
• Existing diagnostic tools may lack accessibility or scalability for widespread farmer use.

Purpose of the Study:

• To introduce a comprehensive, high-resolution dataset of banana leaf images for machine learning applications.
• To facilitate the development of automated systems for detecting Black Sigatoka and Fusarium Wilt Race 1.
• To support agricultural stakeholders in early disease identification and intervention.

Main Methods:

• Collected 11,767 banana leaf images across diverse farms in Tanzania's northern and southern highland regions.
• Categorized images into healthy, Black Sigatoka-affected, and Fusarium Wilt Race 1-affected leaves.
• Ensured unbiased data representation through random farm selection and high-resolution smartphone capture.

Main Results:

• The dataset comprises 3,339 healthy images, 3,496 Black Sigatoka images, and 4,932 Fusarium Wilt Race 1 images.
• The data spans images captured between November 2022 and January 2023.
• The dataset provides a robust foundation for training machine learning models for disease diagnostics.

Conclusions:

• This dataset is a valuable resource for researchers and developers creating AI-driven agricultural solutions.
• Automated disease detection can empower farmers with timely information for crop protection.
• The initiative supports sustainable banana farming through improved disease management strategies.