河南农业大学信息与管理科学学院,郑州 450046
王斌兵,E-mail:wangbinbing@stu.henau.edu.cn
尹飞,E-mail:yin.fei@henau.edu.cn
收稿:2024-11-28,
纸质出版:2025-07-05
移动端阅览
王斌兵,张亚利,郑光,等.EMD-YOLO:基于YOLOv8n改进的轻量化作物叶部病害检测算法[J].华中农业大学学报,2025,44(4):181‐191.
WANG Binbing,ZHANG Yali,ZHENG Guang,et al.EMD-YOLO: an improved lightweight algorithm for detecting diseases in crop leaf based on YOLOv8n[J].Journal of Huazhong Agricultural University,2025,44(04):181-191.
王斌兵,张亚利,郑光,等.EMD-YOLO:基于YOLOv8n改进的轻量化作物叶部病害检测算法[J].华中农业大学学报,2025,44(4):181‐191. DOI: 10.13300/j.cnki.hnlkxb.2025.04.018.
WANG Binbing,ZHANG Yali,ZHENG Guang,et al.EMD-YOLO: an improved lightweight algorithm for detecting diseases in crop leaf based on YOLOv8n[J].Journal of Huazhong Agricultural University,2025,44(04):181-191. DOI: 10.13300/j.cnki.hnlkxb.2025.04.018.
为准确、快速地识别作物叶部病害,降低手工诊断成本,减少叶部病害对作物生长过程与产量的危害,提出一种基于YOLOv8n新型轻量级作物叶部病害检测算法EMD-YOLO。该算法结合多尺度空洞注意力MSDA、EffectiveSE注意力机制、DySample上采样以及Wise-IoU损失函数。其中,多尺度空洞注意力MSDA结合多尺度空间卷积与注意力机制,提高多尺度特征提取效率;EffectiveSE强化特征选择,提升模型表示性能;DySample上采样保留重要特征,提高特征图分辨率和检测性能;Wise-IoU损失函数优化交并比(IoU)计算方式,提升模型定位精度。结果显示,EMD-YOLO的精确度、mAP@0.5、模型权重分别为96.3%、92.8%、4.85 MB,较基线模型YOLOv8n的精确度和平均精度均值分别提高3.0和3.6百分点,权重降低1.4 MB。结果表明,EMD-YOLO的泛化性良好,适用于移动端农作物叶部病害检测设备。
A novel lightweight algorithm for detecting diseases in crop leaf based on YOLOv8n,EMD-YOLO was proposed to accurately and rapidly identify diseases,reduce costs of manual diagnosis,and minimize the impact of leaf diseases on the production and quality of crops.The algorithm integrated the multi-scale dilated attention (MSDA),EffectiveSE attention mechanism,DySample upsampling,and Wise-IoU loss function.MSDA was combined with multi-scale spatial convolution and attention mechanisms to increase the efficiency of extracting multi-scale feature,while EffectiveSE strengthened the selection of feature and improved the representation performance of model.DySample upsampling preserved important features to enhance the map resolution of feature and the detection performance.Wise-IoU loss function optimized intersection over union (IoU) computation to improve the localization accuracy of model.The results showed that the accuracy,mAP@0.5,and weight of EMD-YOLO was 96.3%,92.8%,and 4.85 MB,respectively.The accuracy and mAP@0.5 of YOLOv8n increased by 3.0 and 3.6 percentage points compared with that of the baseline model YOLOv8n,respectively,while the weight decreased by 1.4 MB.It is indicated that EMD-YOLO has good generalization and is suitable for detecting diseases in crop leaf with mobile devices.
ORCHI H,SADIK M,KHALDOUN M,et al.Automation of crop disease detection through conventional machine learning and deep transfer learning approaches[J/OL].Agriculture,2023,13(2):352[2024-11-28].https://doi.org/10.3390/agriculture13020352.
郑果,姜玉松.基于多任务学习农作物叶片病害诊断方法[J].中国农业科技导报,2024,26(1):89-98. ZHENG G,JIANG Y S.Diagnosis of crop disease based on multi-task learning[J].Journal of agricultural science and technology,2024,26(1):89-98 (in Chinese with English abstract) .
彭玉寒,李书琴.基于重参数化MobileNetV2的农作物叶片病害识别模型[J].农业工程学报,2023,39(17):132-140.PENG Y H,LI S Q.Recognizing crop leaf diseases using reparameterized MobileNetV2[J].Transactions of the CSAE,2023,39(17):132-140 (in Chinese with English abstract) .
ROSSI L,KARIMI A,PRATI A.Self-balanced R-CNN for instance segmentation[J/OL].Journal of visual communication and image representation,2022,87:103595[2024-11-28].https://doi.org/10.1016/j.jvcir.2022.103595.
ULLAH A,XIE H M,FAROOQ M O,et al.Pedestrian detection in infrared images using fast RCNN[C]//2018 Eighth International Conference on Image Processing Theory,Tools and Applications (IPTA). November 7-10,2018 ,Xi’an,China.Xi’an:IEEE,2018:1-6.
LIU Y J,ZHOU S J,LIU X Y,et al.Unbiased faster R-CNN for single-source domain generalized object detection[C]//2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). June 16-22,2024 ,Seattle,WA,USA.Seattle:IEEE,2024:28838-28847.
姜晟,曹亚芃,刘梓伊,等.基于改进Faster RCNN的茶叶叶部病害识别[J].华中农业大学学报,2024,43(5):41-50.JIANG S,CAO Y P,LIU Z Y,et al.Recognition of tea leaf disease based on improved Faster RCNN[J].Journal of Huazhong Agricultural University,2024,43(5):41-50 (in Chinese with English abstract) .
WANG L H,SHI W P,TANG Y H,et al.Transfer learning-based lightweight SSD model for detection of pests in Citrus [J/OL].Agronomy,2023,13(7):1710[2024-11-28].https://doi.org/10.3390/agronomy13071710.
ZHU X H,WANG R R,SHI W,et al.Detection of pine-wilt-disease-affected trees based on improved YOLOv7[J/OL].Forests,2024,15(4):691[2024-11-28].https://doi.org/10.3390/f15040691.
OMER S M,GHAFOOR K Z,ASKAR S K.Lightweight improved YOLOv5 model for cucumber leaf disease and pest detection based on deep learning[J].Signal,image and video processing,2024,18(2):1329-1342.
龙阳,肖小玲.基于多注意力机制的苹果叶部病害检测方法[J].江苏农业科学,2023,51(23):178-186.LONG Y,XIAO X L.Apple leaf disease recognition method based on multi-attention mechanism[J].Jiangsu agricultural sciences,2023,51(23):178-186 (in Chinese with English abstract) .
刘春菊,刘延鑫,李斐,等.基于YOLOv5的烟草叶部病害智能识别[J].中国烟草科学,2024,45(3):93-101.LIU C J,LIU Y X,LI F,et al.Intelligent recognition of tobacco leaf diseases based on YOLOv5[J].Chinese tobacco science,2024,45(3):93-101 (in Chinese with English abstract) .
张一帆,张梅,陈杰.基于改进YOLOv7的果蔬叶片病害检测算法[J].兰州工业学院学报,2024,31(4):48-53.ZHANG Y F,ZHANG M,CHEN J.Algorithm for detecting fruit and vegetable leaf diseases based on improved YOLOv7[J].Journal of Lanzhou Institute of Technology,2024,31(4):48-53 (in Chinese with English abstract) .
庞超,王传安,苏煜,等.基于改进YOLOv8n的水稻病害检测方法[J].内蒙古农业大学学报(自然科学版),2024,45(2):62-68.PANG C, WANG C A,SU Y,et al.Rice disease detection method based on improved YOLOv8n[J].Journal of Inner Mongolia Agricultural University (natural science edition),2024,45(2):62-68 (in Chinese with English abstract) .
ZHANG L J,DING G C,LI C R,et al.DCF-YOLOv8:an improved algorithm for aggregating low-level features to detect agricultural pests and diseases[J/OL].Agronomy,2023,13(8):2012[2024-11-28].https://doi.org/10.3390/agronomy13082012.
KAGGLE.Tomato[DB/OL].[2024-11-28].https://www.kaggle.com/datasets/noulam/tomato.
KAGGLE.Corn or maize leaf disease dataset[DB/OL].[2024-11-28].https://www.kaggle.com/datasets/smaranjitghose/corn-or-maize-leaf-disease-dataset.
左昊轩,黄祺成,杨佳昊,等.基于双目视觉和改进YOLOv8的玉米茎秆宽度原位识别方法[J].智慧农业,2023,5(3):86-95.ZUO H X,HUANG Q C,YANG J H,et al. In situ identification method of maize stalk width based on binocular vision and improved YOLOv8[J].Smart agriculture,2023,5(3):86-95 (in Chinese with English abstract) .
YUE X,QI K,NA X Y,et al.Improved YOLOv8-seg network for instance segmentation of healthy and diseased tomato plants in the growth stage[J/OL].Agriculture,2023,13(8):1643[2024-11-28]. https://doi.org/10.3390/agriculture13081643.
孙向阳,杨晓霞.GSW-YOLOv7:一种基于改进YOLOv7的玉米叶病害检测方法[J].山东农业大学学报(自然科学版),2024,55(4):566-578.SUN X Y,YANG X X.GSW-YOLOv7:a maize leaf disease detection method based on improved YOLOv7[J].Journal of Shandong Agricultural University (natural science edition),2024,55(4):566-578 (in Chinese with English abstract) .
LIU S,QI L,QIN H F,et al.Path aggregation network for instance segmentation[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recogni tion. June 18-23,2018 ,Salt Lake City,UT,USA.Salt Lake City:IEEE,2018:8759-8768.
WANG X W,LIU J.Vegetable disease detection using an improved YOLOv8 algorithm in the greenhouse plant environment[J/OL].Scientific reports,2024,14:4261[2024-11-28].https://doi.org/10.1038/s41598-024-54540-9.
岳凯,张鹏超,王磊,等.基于改进YOLOv8n的复杂环境下柑橘识别[J].农业工程学报,2024,40(8):152-158.YUE K,ZHANG P C,WANG L,et al.Recognizing citrus in complex environment using improved YOLOv8n[J].Transactions of the CSAE,2024,40(8):152-158 (in Chinese with English abstract) .
王帅, 刘珊珊, 李保田, 等.基于改进YOLOv8s的织物疵点检测[J].棉纺织技术,2025,53(2): 49-55.WANG S, LIU S S, LI B T, et al.Fabric defect detection based on improved YOLOv8s[J].Cotton textile technology,2025,53(2): 49-55 (in Chinese with English abstract) .
JIAO J Y,TANG Y M,LIN K Y,et al.DilateFormer:multi-scale dilated transformer for visual recognition[DB/OL].arXiv,2023:2302.01791[2024-11-28].https://arxiv.org/abs/2302.01791v1.
LEE Y,PARK J.CenterMask:real-time anchor-free instance segmentation[DB/OL].arXiv, 2019 :1911.06667[2024-11-28].https://arxiv.org/abs/1911.06667v6.
LIU W Z,LU H,FU H T,et al.Learning to upsample by learning to sample[DB/OL].arXiv,2023:2308.15085[2024-11-28].https://arxiv.org/abs/2308.15085v1.
TONG Z J,CHEN Y H,XU Z W,et al.Wise-IoU:bounding box regression loss with dynamic focusing mechanism[DB/OL].arXiv,2023:2301.10051[2024-11-28].https://arxiv.org/abs/2301.10051v3.
SAITO K,USHIKU Y,HARADA T,et al.Strong-weak distribution alignment for adaptive object detection[DB/OL].arXiv,2018:1812.04798[2024-11-28].https://arxiv.org/abs/1812.04798v3.
HU J,SHEN L,ALBANIE S,et al.Squeeze-and-excitation networks[DB/OL].arXiv,2017:1709.01507[2024-11-28].https://arxiv.org/abs/1709.01507v4.
WANG J Q,CHEN K,XU R,et al.CARAFE:content-aware ReAssembly of FEatures[DB/OL].arXiv, 2019 :1905.02188[2024-11-28].https://arxiv.org/abs/1905.02188v3.
LU H,LIU W Z,FU H T,et al.FADE:fusing the assets of decoder and encoder for task-agnostic upsampling[DB/OL].arXiv,2022:2207.10392[2024-11-28].https://arxiv.org/abs/2207.10392v2.
LU H,LIU W Z,YE Z X,et al.SAPA:similarity-aware point affiliation for feature upsampling[DB/OL].arXiv,2022:2209.12866[2024-11-28].https://arxiv.org/abs/2209.12866v2.
ZHANG Y F,REN W Q,ZHANG Z,et al.Focal and efficient IoU loss for accurate bounding box regression[J].Neurocomputing,2022,506:146-157.
KAGGLE.New Plant Diseases Dataset[DB/OL].[2024-11-28].https://www.kaggle.com/datasets/vipoooool/new-plant-diseases-dataset.
0
浏览量
2
下载量
2
CSCD
关联资源
相关文章
相关作者
相关机构
京公网安备11010802024621