Comparative Performance Analysis of YOLOv10 and YOLOv11 for Automated Mulberry Leaf Nutrient Deficiency Detection

Early identification of nutrient deficiencies in mulberry leaves is essential for maintaining leaf quality,improving silkworm productivity, and supporting sustainable sericulture practices. Conventional assessment methodsrely heavily on manual observation, which can be time-consuming, subjective, and prone to inconsistencies. Recentadvances in deep learning-based object detection have enabled automated and accurate analysis of plant health conditionsfrom digital images. This study presents a comparative performance analysis of YOLOv10 and YOLOv11 for automatedmulberry leaf nutrient deficiency detection. A dataset comprising approximately 6,000 annotated images representing sixclasses, namely Healthy, Nitrogen Deficiency, Potassium Deficiency, Phosphorus Deficiency, Iron Deficiency, and SulphurDeficiency, was utilized for model training and evaluation. The dataset was divided into training, validation, and testingsubsets using a 70:15:15 ratio. Standard preprocessing and data augmentation techniques were applied to improve modelgeneralization and robustness.