Abstract:Aiming at the problem of low detection accuracy due to the different scale sizes of apple leaf disease spots and their similarity to the background, this paper proposes a multi-scale detection network called MSL-Net. Firstly, a multiplexed aggregated feature extraction network is proposed using RES-Bottleneck and middlepartial-convolution (MP-Conv) to capture multi-scale spatial features and enhance focus on disease features for better differentiation between disease targets and background information. Secondly, a lightweight feature fusion network is designed using scale fuse concat (SF-Cat) and triple scale sequence feature fusion (TSSF) module to merge multi-scale feature maps comprehensively. DWConv and GhostNet lighten the network, while C3-GN reduces missed detections by suppressing irrelevant background information. Finally, Soft-NMS is used in the post-processing stage to improve the problem of misdetection of dense disease sites. The results show that MSL-Net improves mAP@0.5 by 2.0% over the baseline YOLOv5s and reduces parameters by 44%, reducing computation by 27%, outperforming other SOTA models overall. This method also shows excellent performance compared to the latest research.