This study presents the development of Fog-Powered CropGuard: Live Health Monitoring System for Tomato, Corn, and Bell Pepper Using YOLOv8n and VGG-16, designed to detect and monitor plant health for tomato, corn, and bell pepper. The system integrates Artificial Intelligence (AI), Internet of Things (IoT), and fog computing to provide real-time plant disease detection with reduced dependency on internet connectivity. A USB camera is used to capture live images of plant leaves, while a Raspberry Pi serves as the fog device that processes the data locally. The system utilizes a fine-tuned YOLOv8n model for plant leaf detection and crop identification, and a pre-trained VGG-16 model for classifying leaf conditions as healthy or diseased. The results are processed to determine the overall plant health status based on the ratio of diseased leaves. A web-based interface is developed to display real-time monitoring results, including alerts and recommended actions. The system was evaluated based on detection performance, classification accuracy, responsiveness, and reliability. The YOLOv8n model achieved a precision of 91.23%, recall of 87.91%, [email protected] of 91.79%, and [email protected]:0.95 of 83.76%, indicating strong performance in plant leaf detection and crop identification. Meanwhile, the VGG-16 model achieved an accuracy of 84.00%, precision of 85.42%, recall of 82.00%, and F1-score of 83.67%, showing consistent classification performance in identifying healthy and diseased leaves. The system also achieved an average inference latency of 319.60 ms per image, making it suitable for near real-time monitoring. The findings indicate that combining AI, IoT, and fog computing provides a practical and efficient solution for smart agriculture, especially in areas with limited internet connectivity.
Fog computing, plant disease detection, YOLOv8n, VGG-16, Internet of Things (IoT), real-time monitoring, smart agriculture