基于YOLOv5s-GCB模型的苹果叶部病害检测研究

doi:10.12190/j.issn.2096-1197.2025.04.11

摘要/Abstract

摘要： 【目的】提出一种基于YOLOv5s-GCB模型的苹果叶部病害检测方法,旨在实现模型轻量化的同时提升检测精度。【方法】在YOLOv5s框架基础上,引入Ghost卷积以减少卷积计算量,实现模型轻量化;在 Neck 部分嵌入坐标CA注意力机制,增强对苹果叶片病斑区域的特征关注能力;并采用双向加权特征金字塔结构（BiFPN）优化多尺度病斑特征融合。基于采集的苹果叶部病害图像数据集对改进模型YOLOv5s-GCB进行训练与测试。【结果】改进的YOLOv5s-GCB模型在苹果叶部病害检测任务中表现优异,精确率（Precision）、召回率（Recall）与平均精度均值（mAP@0.5）分别达到 90.7%、87.4% 和 93.4%。在苹果叶部斑点落叶病、灰斑病、锈病的检测中,YOLOv5s-GCB模型的mAP@0.5均最高,分别为93.8%、93.4%、93.0%。【结论】改进的YOLOv5s-GCB模型不仅具备高精度和高速度的检测能力,且模型小,适用于苹果叶部病害的自动化智能识别,具有较强的实用价值。

关键词: YOLOv5s, Ghost, CA注意力机制, 双向加权特征金字塔结构（BiFPN）, 苹果叶部病害, 检测

Abstract: 【Objective】To propose an apple leaf disease detection method based on the YOLOv5s-GCB model,aiming to achieve a lightweight model while improving detection accuracy. 【Methods】On the basis of the YOLOv5s framework,Ghost convolution was introduced to reduce the computational load of convolution operations,thereby achieving a model lightweight;a coordinate attention（CA） mechanism was embedded in the Neck component to enhance the feature attention capability toward diseased spot regions on apple leaves;and a bidirectional weighted feature pyramid network（BiFPN） was adopted to optimize the multi-scale fusion of diseased spot features. The improved YOLOv5s-GCB model was trained and tested on a collected dataset of apple leaf disease images. 【Results】The improved YOLOv5s-GCB model demonstrates excellent performance in apple leaf disease detection tasks,with Precision,Recall,and mean Average Precision（mAP@0.5）reaching 90.7%,87.4%,and 93.4%,respectively. For the detection of Apple Alternaria blotch,gray leaf spot disease,and rust disease,the YOLOv5s-GCB model achieves the highest mAP@0.5 values of 93.8%,93.4%,and 93.0%,respectively. 【Conclusion】The improved YOLOv5s-GCB model not only possesses high-precision and high-speed detection capabilities but also features a compact model size,making it suitable for automated intelligent recognition of apple leaf diseases and demonstrating strong practical value.

Key words: YOLOv5s, Ghost, Coordinate attention（CA） mechanism, Bidirectional weighted feature pyramid network（BiFPN）, Apple leaf diseases, Detection

中图分类号:

S436.611.1

赵兴, 王迎超, 刘纪博. 基于YOLOv5s-GCB模型的苹果叶部病害检测研究[J]. 北方农业学报, 2025, 53(4): 121-134.

ZHAO Xing, WANG Yingchao, LIU Jibo. Research on apple leaf disease detection based on the YOLOv5s-GCB model[J]. Journal of Northern Agriculture, 2025, 53(4): 121-134.

参考文献

[1] SRINIVASAN K,DASTOOR P H,RADHAKRISHNAIAH P,et al.FDAS:a knowledge-based framework for analysis of defects in woven textile structures[J].The Journal of the Textile Institute,1992,83(3):431-448.
[2] 曾晏林,贺壹婷,蔺瑶,等.基于BCE-YOLOv5的苹果叶部病害检测方法[J].江苏农业科学,2023,51(15):155-163.
[3] 周绍发,肖小玲,刘忠意,等.改进的基于YOLOv5s苹果树叶病害检测[J].江苏农业科学,2023,51(13):212-220.
[4] 刘洋,冯全,王书志,等.基于轻量级CNN的植物病害识别方法及移动端应用[J].农业工程学报,2019,35(17):194-204.
[5] 张俊宁,毕泽洋,闫英,等.基于注意力机制与改进YOLO的温室番茄快速识别[J].农业机械学报,2023,54(5):236-243.
[6] 张小花,李浩林,李善军,等.基于EfficientDet-D1的草莓快速检测及分类[J].华中农业大学学报,2022,41(6):262-269.
[7] 袁杰,谢霖伟,郭旭,等.基于改进YOLOv7的苹果叶片病害检测方法[J].农业机械学报,2024,55(11):68-74.
[8] HAN K,WANG Y H,TIAN Q,et al.GhostNet:more features from cheap operations[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR).Seattle:IEEE,2020:1577-1586.
[9] SALEEM M H,POTGIETER J,ARIF K M.A performance-optimized deep learning-based plant disease detection approach for horticultural crops of New Zealand[J].IEEE Access,2022,10:89798-89822.
[10] SUN Y Y,JIANG Z H,ZHANG L P,et al.SLIC—SVM based leaf diseases saliency map extraction of tea plant[J].Computers and Electronics in Agriculture,2019,157:102-109.
[11] BISWAS S,JAGYASI B,SINGH B P,et al.Severity identification of potato late blight disease from crop images captured under uncontrolled environment[C]//2014 IEEE Canada International Humanitarian Technology Conference-(IHTC).Montreal:IEEE,2014:1-5.
[12] KHITTHUK C,SRIKAEW A,ATTAKITMONGCOL K,et al.Plant leaf disease diagnosis from color imagery using co-occurrence matrix and artificial intelligence system[C]//2018 International Electrical Engineering Congress(iEECON).Krabi:IEEE,2018:1-4.
[13] SETHY P K,BARPANDA N K,RATH A K,et al.Deep feature based rice leaf disease identification using support vector machine[J].Computers and Electronics in Agriculture,2020,175:105527.
[14] SINGH S,GUPTA S,TANTA A,et al.Extraction of multiple diseases in apple leaf using machine learning[J].International Journal of Image and Graphics,2022,22(3):2140009.
[15] ALI S,HASSAN M,KIM J Y,et al.FF-PCA-LDA:intelligent feature fusion based PCA-LDA classification system for plant leaf diseases[J].Applied Sciences,2022,12(7):3514.
[16] HASAN R I,YUSUF S M,ALZUBAIDI L.Review of the state of the art of deep learning for plant diseases:a broad analysis and discussion[J].Plants,2020,9(10):1302.
[17] REN S Q,HE K M,GIRSHICK R,et al.Faster R-CNN:towards real-time object detection with region proposal networks[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2017,39(6):1137-1149.
[18] LIU B,ZHAO W C,SUN Q Q.Study of object detection based on faster R-CNN[C]//2017 Chinese Automation Congress(CAC).Jinan:IEEE,2017:6233-6236.
[19] USHADEVI G,GOKULNATH B V.A survey on plant disease prediction using machine learning and deep learning techniques[J].Inteligencia Artificial,2020,23(65):136-154.
[20] SUN H,XU H,LIU B,et al.MEAN-SSD:a novel real-time detector for apple leaf diseases using improved light-weight convolutional neural networks[J].Computers and Electronics in Agriculture,2021,189:106379.
[21] OBSIE E Y,QU H C,ZHANG Y J,et al.YOLOv5s-CA:an improved YOLOv5 based on the attention mechanism for mummy berry disease detection[J].Agriculture,2023,13(1):78.
[22] REHMAN S,ATTIQUE KHAN M,ALHAISONI M,et al.A framework of deep optimal features selection for apple leaf diseases recognition[J].Computers,Materials and Continua,2023,75(1):697-714.
[23] KITSIRANUWAT S,KAWICHAI T,KHANARSA P.Identification and classification of diseases based on object detection and majority voting of bounding boxes[J].Journal of Advances in Information Technology,2023,14(6):1301-1311.
[24] 叶东东,徐霞,丁玉庭.深度学习与机器视觉在鱼类加工与品质监测中的研究进展[J/OL].食品与发酵工业,2024:389-398.(2024-03-21).https://link.cnki.net/doi/10.13995/j.cnki.11-1802/ts.038903.
[25] 李键,李华,胡翔坤,等.基于深度学习的表面缺陷检测技术研究进展[J].计算机集成制造系统,2024,30(3):774-790.
[26] 夏玉果,董天天,丁晟.基于轻量化深度迁移神经网络的电子元器件识别[J].电子器件,2023,46(6):1673-1679.
[27] 贺苏强. 基于显微高光谱成像技术的癌细胞鉴别方法研究[D].西安:西安石油大学,2023.
[28] 蔡舒妤,殷航,史涛,等.基于ResNet-LSTM的航空发动机性能异常检测方法[J].航空发动机,2024,50(1):135-142.
[29] 杜浩飞,张超,李建军.基于SENet-ResNext-LSTM的风机轴承故障诊断[J].机械强度,2023,45(6):1271-1279.
[30] 曹义亲,刘文才,徐露.基于 YOLOX-αSMV的带钢材料表面缺陷检测算法[J/OL].华东交通大学学报,2024:109-117.(2024-03-28).https://link.cnki.net/doi/10.16749/j.cnki.jecjtu.20240326.010.
[31] WOO S,PARK J,LEE J Y,et al.CBAM:convolutional block attention module[C]//Proceedings of the 15th European Conference on Computer Vision(ECCV). Germany:Springer,2018:3-19.
[32] ZHANG S,CHE S B,LIU Z,et al.A real-time and lightweight traffic sign detection method based on ghost-YOLO[J].Multimedia Tools and Applications,2023,82(17):26063-26087.
[33] 高明. 基于进化算法的卷积神经网络图像去噪算法研究[D].南京:南京邮电大学,2023.
[34] 郭惠萍,曹亚州,王晨思,等.基于迁移学习的苹果落叶病识别与应用[J/OL].农业工程学报,2024:1-9.(2024-03-06).https://kns.cnki.net/kcms/detail/11.2047.S.20240301.1738.030.html.
[35] 宋中山,肖博文,艾勇,等.基于改进YOLOv4的轻量化目标检测算法[J].电子测量技术,2022,45(16):142-152.
[36] QI C,GAO J F,PEARSON S,et al.Tea Chrysanthemum detection under unstructured environments using the TC-YOLO model[J].Expert Systems with Applications,2022,193:116473.
[37] TAN M,PANG R,LE Q V.EfficientDet:scalable and efficient object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Publisher:IEEE,2020.
[38] 张会敏,谢泽奇.基于多尺度注意力卷积神经网络的苹果叶部病害识别方法[J].江苏农业科学,2023,51(16):154-161.