Cauliflower leaf diseases: A computer vision dataset for smart agriculture

Area of Science:

• Agricultural Science
• Computer Vision
• Machine Learning

Background:

• Cauliflower leaf diseases are difficult to diagnose early, leading to rapid spread and significant crop losses.
• Current disease management often relies on extensive pesticide use, which is costly and environmentally harmful.
• Accurate and timely disease diagnosis is crucial for effective crop management and sustainable agriculture.

Purpose of the Study:

• To introduce a comprehensive dataset of cauliflower leaf images to accelerate research in plant disease detection.
• To facilitate the development of advanced machine learning models for early identification and monitoring of cauliflower diseases.
• To support the advancement of precision agriculture techniques through enhanced disease diagnosis.

Main Methods:

• Collected 2,661 images of cauliflower leaves across diverse locations, weather conditions, and capture devices in Bangladesh.
• Categorized images into three classes: Healthy, Insect Holes, and Black Rot.
• Preprocessed images by resizing (3000x3000 pixels), adjusting brightness, and removing duplicates/low-quality samples to ensure training readiness.

Main Results:

• A high-quality, diverse dataset ready for training deep learning models for cauliflower leaf disease identification.
• The dataset enables the development of accurate models for early disease detection, surpassing manual diagnosis limitations.
• Potential for real-time applications, including mobile apps and smart farming equipment integration for immediate farmer action.

Conclusions:

• The introduced dataset is a valuable resource for agricultural research, particularly in precision agriculture and disease management.
• It will significantly contribute to developing highly accurate machine learning models for automated monitoring and smart decision-making.
• The dataset supports practical, real-time disease identification, empowering farmers with actionable insights and reducing reliance on expert knowledge.