<![CDATA[This study addresses the inaccurate detection of mango (Mangifera indica) leaf diseases, which can reduce plant productivity. A deep-learning-based automatic detection system is proposed to identify five leaf diseases (Mangifera indica): anthracnose, cutting weevil, dieback, powdery mildew, and sooty mold. This study compares two object detection models, namely YOLOv12n and YOLOv26n, with a dataset of 1,970 images. The data is annotated and divided into training, validation, and testing with weights of 70%, 20%, and 10%, respectively. Both YOLO models were trained for 100 epochs and evaluated using accuracy, precision, recall, and F1 score. The results of this study indicate that YOLOv26n performs better during testing, with an average accuracy of 97,06%, a recall of 98,76%, a precision of 98,25%, and an F1 score of 98,50%. In comparison, YOLOv12n achieved 94,52% accuracy, 98,92% recall, 95,47% precision, and 97,16% F1 score. Although YOLOv12n had slightly better training loss, YOLOv26n delivered more consistent performance, particularly in mean average precision (mAP) at higher thresholds. Therefore, YOLOv26n is better at identifying mango leaf diseases and has greater potential for real-time agricultural applications.]]>
Copyrights © 2026
