YOLO-ODD: An Improved YOLOv8s Model for Onion Foliar Disease Detection

Key findings

• ▸The best-performing variant — YOLOv8 with a Dynamic Task-Aligned Head (DTAH) — reached 77.3% mAP50, 81.5% precision, and 72.1% recall, improving on baseline YOLOv8s (72.9% mAP50) and YOLOv5s (65.4% mAP50).
• ▸It also beat two much larger two-stage detectors, RetinaNet-R50 (70.34% mAP50, 36M parameters, 65 FPS) and Faster R-CNN-R101 (72.05% mAP50, 44M parameters, 44 FPS), while running at 123 FPS with only 11.1M parameters.
• ▸Adding a second attention mechanism (CBAM) on top of DTAH did not improve results in this trial — the combined CBAM+DTAH model scored lower (75.0% mAP50) than DTAH alone, suggesting DTAH's deformable, task-aligned head was already well-matched to irregular disease-lesion shapes.
• ▸Twister disease was the easiest of the four diseases to detect across every model tested (up to 85.5% mAP50), while Stemphylium blight was consistently the hardest, often confused with purple blotch or background due to visual similarity between the two diseases.

Abstract

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. anthracnose, Stemphylium blight, purple blotch (PB), and Twister disease. The experimental results demonstrate that customized 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% precision and Recall of 72.10% and thus this 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.

Read the story