Lightweight Apple Leaf Disease Detection Based on YOLOv5
To improve the accuracy of apple leaf disease recognition in practical applications,a YOLOv5 based apple leaf disease recognition model is proposed.To ensure the smooth operation of the model on actual orchard mobile devices,the Ghost module and Ghost BottleNeck structure are introduced in YOLOv5 for lightweighting.Replace the loss function with Alpha-IOU loss function to improve the detection accuracy of the model.Verified on the apple leaf disease dataset,compared with the original YOLOv5 algorithm,based on the improved YOLOv5 algorithm mAP@0.5 increased by 1.3%,mAP@0.5:0.95 increased by 1.4%,while the parameter count decreased by 5.2 GFLOPs.The results indicate that the proposed improved YOLOv5 algorithm can better meet the requirements of apple leaf disease detection in practical situations.
NETL
NSTL
万方数据
