Onion foliar disease detection using YOLOv8 deep learning model YOLO-ODD: an improved YOLOv8s model for onion foliar disease detection

Anusha Raj¹Mukund Dawale¹Sagar Wayal¹Kiran Khandagale¹Indira Bhangare¹Susmita Banerjee²Ashwini Gajarushi²Rajbabu Velmurugan³Maryam Shojaei Baghini³Suresh Gawande^1*

• ¹Directorate of Onion and Garlic Research (ICAR), Pune, India
• ²TIH-IIT Mumbai, Mumbai, Maharashtra, India
• ³Indian Institute of Technology Bombay, Mumbai, Maharashtra, India

Onion crops are affected by many diseases at different stages of growth, resulting in significant yield loss. The early detection of diseases helps in the timely incorporation of management practices, thereby reducing yield losses. However, the manual identification of plant diseases requires considerable effort and is prone to mistakes. Thus, adopting cutting-edge technologies such as machine learning (ML) and deep learning (DL) can help overcome these difficulties by enabling the early detection of plant diseases. This study presents a cross layer integration of YOLOv8 architecture for detection of onion leaf diseases viz.This study aims to provide an artificial intelligence-based solution for the detection of onion diseases (anthracnose, Stemphylium blight, purple blotch (PB), and Twister disease.) by training the different deep learning algorithms viz; YOLOv5, YOLOv8, Faster R-CNN and RetinaNet. Among these, YOLOv5 performed best with an mAP50 of 80.8, making it the most reliable model for detecting onion diseases followed by the YOLOv8 with mAP50 of 77.1. Further, class wise performance of YOLO based models found superior over the Faster R-CNN and RetinaNet models. The experimental results demonstrate that customised YOLOv8 model YOLO-ODD integrated with CABM and DTAH attentions outperform YOLOv5 and YOLO v8 base models in most disease categories, particularly in detecting Anthracnose, Purple Blotch, and Twister disease. Proposed YOLOv8 model achieved the highest overall 77.30% accuracy, 81.50% precession and Recall of 72.10%…. and thus YOLOv8-based deep learning approach will detect and classify major onion foliar diseases while optimizing for accuracy, real-time application, and adaptability in diverse field conditions.The detection and identification results for the YOLOv8 approach were validated by prominent statistical metrics, such as detection precision, recall, mAP@50 value, and F1-score, which were 78.3%, 71.5%, 73.7%, and 74.6%, respectively. The experimental results demonstrate that using YOLOv85 algorithm helps in the identification of onion plant diseases and in understanding disease severity, thus encouraging us to consider YOLOv85 as a baseline model.