Smart Agriculture: Deep Learning-Based Plant Disease Diagnosis with YOLOv8 and VGG19

Plant diseases, causing up to 40% an- nual crop losses, threaten global food security and economic stability. Traditional diagnostics, such as visual inspection and laboratory testing, are sub- jective, slow, and unscalable. LeafScan integrates YOLOv8, YOLOv9, a custom CNN, and VGG19 to deliver early plant disease detection with 99.0% accuracy (YOLOv9) and 98.4% [email protected]. Trained on a 147,500-image dataset with advanced augmenta- tion, LeafScan outperforms state-of-the-art models via precise localization and robust classification. Deployed through a mobile application optimized for low-end devices, it enables real-time diagnostics, reducing crop losses by 30% and pesticide use by 20%. Extensive testing across diverse crops, dis- eases, and environmental conditions validates its scalability and robustness, supporting sustainable agriculture and SDG 2 (Zero Hunger).