基于DBS-YOLOv11n模型的苹果叶片病害检测方法OA

To address the problems of poor adaptability to lesion morphology,insufficient synergy between global and local features,and difficulty in detecting small-scale lesions in apple leaf disease detection,this paper proposed an apple leaf disease detection method based on the DBS-YOLOv11n model.First,DynamicConv was introduced into the C3K2 structure in the backbone network and neck network to enhance the recognition capability of randomly distributed and morphologically variable features of apple leaf lesions through dynamic convolution kernel combinations.Second,the BRA(Bi-level Routing Attention)module was used to reconstruct C2PSA,achieving collaborative perception between overall leaf condition assessment and local lesion identification through a bi-level routing attention mechanism.Finally,the SEAM(Separated and Enhancement Attention Module)was integrated into the detection head to improve the model's detection capability and feature representation for small-scale lesions through attention enhancement mechanisms.Based on the detection results,corresponding prevention and control recommendations were provided for different disease types.Experimental results showed that the improved DBS-YOLOv11n model achieved improvements of 3.9,2.9,and 3.6 percentage points in precision,recall,and mAP@0.5 compared with the YOLOv11n model,reaching as 95.0％,85.3％,and 91.4％.Compared with classical object detection models,the mAP was improved by an average of 9.4％.This study significantly improved the accuracy of apple leaf disease detection,providing technical support for intelligent orchard management.