AI-based Plant Disease Detection: Development and Deployment of YOLO Models for Agricultural Innovation

This research focuses on enhancing plant disease detection by developing four models based on YOLOv11, the latest iteration in the “You Only Look Once” series, renowned for real-time object detection capabilities. A comparative analysis is conducted with the open-source YOLOX model and the YOLOvME model to evaluate performance metrics such as accuracy, precision, recall, and mean Average Precision. Notably, the Apple disease detection model outperforms the others, achieving an impressive accuracy of $\mathbf{9 3. 8 \%}$. The models are trained using a comprehensive dataset comprising images of various plant diseases, enabling the identification and classification of multiple disease types. The deployment of these models on ALIVEculture.ca (a dedicated application for plant disease detection) allows users to perform real-time analyses through multiple imaging devices, including smartphone cameras and remote monitoring systems in greenhouses. Additionally, we develop a mobile application that leverages these AI models, providing on-the-go disease detection and immediate feedback.