AI based real time disease diagnosis in plants using deep learning driven CNNs.

Area of Science:

• Agricultural Science
• Computer Science
• Artificial Intelligence

Background:

• Traditional plant disease diagnosis is slow, labor-intensive, and prone to human error.
• Current methods are unsuitable for large-scale crop systems requiring real-time, accurate diagnostics.
• Early detection is crucial for maximizing crop yield and minimizing losses.

Purpose of the Study:

• To develop and validate a real-time plant disease diagnosis system using deep learning.
• To improve the speed, accuracy, and scalability of plant disease detection.
• To support precision agriculture and sustainable plant health management.

Main Methods:

• Implementation of a Plant Disease Diagnosis using Deep Learning (PDD-DL) framework.
• Utilizing Convolutional Neural Networks (CNNs) for automated analysis of plant images.
• Model validation on common crops with potential for retraining on diverse disease classes.

Main Results:

• The PDD-DL model achieved high performance metrics: 98.32% accuracy, 97.85% precision, 98.14% recall, and 97.99% F1-score.
• Real-time inference speed of 42.6 ms per image demonstrates system efficiency.
• The model effectively differentiates between healthy and diseased plants.

Conclusions:

• Deep learning, specifically CNNs, offers a faster, more trustworthy, and scalable alternative to traditional plant disease diagnosis.
• The developed framework enhances accuracy and speed in diagnosing plant diseases.
• This technology aids in precision agriculture and promotes sustainable crop management practices.