Wheat Powdery Mildew Automatic Identification Through YOLOv8 Instance Segmentation

The global production of wheat is seriously endangered by wheat powdery mildew, which can be brought on by Blumeria graminis f. sp. tritici (Bgt). The illness produces large annual losses and threatens food security. The labourintensive and time-consuming characteristics of conventional detection methods emphasizes the need for elegant, automated surveillance systems. Deep learning techniques, particularly the YOLOv8 instance segmentation model, were employed in this study to enhance wheat powdery mildew detection and classification. Results indicate the YOLOv8 model is efficient; it can recognize wheat leaves affected by powdery mildew with a precision of 99.37%, recall of 96%, and an F1-score of 97.67%. This approach works superior to before methods, making it a promising solution for early detection of illnesses in wheat crops. In addition, the research's next steps are examined, highlighting the potential of improving the YOLOv8 model in an array of environmental situations and disease stages. Effective disease detection could be enhanced by merging new data sources as drone-based monitoring systems and hyperspectral imaging. Developing and setting up autonomous disease tracking systems in real agricultural settings involves collaboration among data scientists, technologists, and agronomic teams, in addition to user-friendly interfaces tailored to farmer needs. When everything is taken into account, this study promotes the development of successful tools to control diseases, promoting sustainable crop production and assuring food safety around the globe.