Multi-Plant Leaf Disease Detection Using AlexNet and MobileNet CNN Methods

Timely and accurate detection of plant diseases is crucial for ensuring crop health, minimizing agricultural losses, and enhancing food security. Traditional disease identification methods, such as manual inspection, are inefficient, labourintensive, and impractical for large-scale farming. This study leverages deep learning-based approaches using AlexNet for binary classification (healthy vs. diseased plants) and MobileNet for multi-class classification (identifying specific plant diseases). A custom dataset of 10,000 images, categorized into healthy and diseased leaves across multiple plant species, is used for model training and evaluation. Preprocessing techniques, including image augmentation and normalization, enhance model robustness. The models are assessed based on accuracy, precision, recall, and F1-score, with MobileNet demonstrating efficiency in real-time applications and AlexNet excelling in disease detection at a binary level. Additionally, a benchmark comparison with ResNet and EfficientNet is conducted to analyse trade-offs in accuracy and computational efficiency. The findings support AI-driven disease detection for mobile and IoT-based agricultural systems, particularly benefiting farmers in resourcelimited environments. Future work will focus on model optimization for real-world deployment and integration into smart agricultural monitoring systems.