|本期目录/Table of Contents|

[1]刘宇雄,兰艳亭,陈晓栋.基于级联式分组注意力机制的葡萄病害识别模型[J].江苏农业科学,2025,53(5):121-128.
 Liu Yuxiong,et al.Grape disease recognition model based on cascaded group attention mechanism[J].Jiangsu Agricultural Sciences,2025,53(5):121-128.
点击复制

基于级联式分组注意力机制的葡萄病害识别模型(PDF)
分享到:

《江苏农业科学》[ISSN:1002-1302/CN:32-1214/S]

卷:
第53卷
期数:
2025年第5期
页码:
121-128
栏目:
病害智能检测
出版日期:
2025-03-05

文章信息/Info

Title:
Grape disease recognition model based on cascaded group attention mechanism
作者:
刘宇雄1兰艳亭1陈晓栋2
1.中北大学电气与控制工程学院,山西太原 030000; 2.北京中农博后农业科学研究院/中国健康农业产学研协同创新平台,北京 100193
Author(s):
Liu Yuxionget al
关键词:
葡萄病害模型鲁棒性视觉Transformer图像识别
Keywords:
-
分类号:
S126;TP391.41
DOI:
-
文献标志码:
A
摘要:
在农业应用场景中,天气和环境因素对图像质量的影响要求机器视觉模型具有较好的鲁棒性。强鲁棒性的模型一般具有更高的浮点计算量和参数量,不利于模型的部署。针对此问题,提出一种基于级联式分组注意力机制的葡萄病害识别算法,以轻量型Transformer架构EfficientVit为基础,将其中的前向传播网络以卷积门控线性单元代替,在减少参数量的同时增加模型的鲁棒性;同时利用部分卷积代替深度卷积进行相对位置信息的编码,在不影响模型计算复杂度的情况下提高模型的检测精度。经测试,本研究所提出的模型在拥有6种病害的自然环境下的葡萄图像数据集上实现了98%的检测准确率。消融试验结果表明,与原始EfficientVit模型相比,改进的模型在降低16.3%浮点计算量(0.74 G)和16.9%参数量(12.83 M)的条件下使模型的检测准确率提高3.9百分点。在利用大气散射模型加入雾效果噪声后验证集上实现了最佳的鲁棒性,在调整亮度后的验证集上的鲁棒性表现仅低于参数量为所提模型2倍的Swin Transformer-Tiny(Swin-Tiny)。本研究提出的模型以较低的模型计算复杂度和参数量,实现了高检测准确率,同时在4种不同噪声验证集中实现不低于76%的分类准确率,模型鲁棒性表现优异,更复合实际农业应用场景的要求。
Abstract:
-

参考文献/References:

[1]Fraiwan M,Faouri E,Khasawneh N. Multiclass classification of grape diseases using deep artificial intelligence[J]. Agriculture,2022,12(10):1542.
[2]Li Z,Paul R,Tis T B,et al. Non-invasive plant disease diagnostics enabled by smartphone-based fingerprinting of leaf volatiles[J]. Nature Plants,2019,5(8):856-866.
[3]Savary S,Ficke A,Aubertot J N,et al. Crop losses due to diseases and their implications for global food production losses and food security[J]. Food Security,2012,4(4):519-537.
[4]Faithpraise F,Birch P,Young R,et al. Automatic plant pest detection and recognition using k-means clustering algorithm and correspondence filters[J]. International Journal of Advanced Biotechnology and Research,2013,4(2):189-199.
[5]Arnal Barbedo J G. Digital image processing techniques for detecting,quantifying and classifying plant diseases[J]. SpringerPlus,2013,2(1):660.
[6]Zhang F,Chen Z J,Ali S,et al. Multi-class detection of cherry tomatoes using improved YOLO v4-Tiny[J]. International Journal of Agricultural and Biological Engineering,2023,16(2):225-231.
[7]彭红星,徐慧明,刘华鼐. 融合双分支特征和注意力机制的葡萄病虫害识别模型[J]. 农业工程学报,2022,38(10):156-165.
[8]Zhang P,Yang L,Li D L. EfficientNet-B4-ranger:a novel method for greenhouse cucumber disease recognition under natural complex environment[J]. Computers and Electronics in Agriculture,2020,176:105652.
[9]Guo W J,Feng Q,Li X Z,et al. Grape leaf disease detection based on attention mechanisms[J]. International Journal of Agricultural and Biological Engineering,2022,15(5):205-212.
[10]王瑞鹏,陈锋军,朱学岩,等. 采用改进的EfficientNet识别苹果叶片病害[J]. 农业工程学报,2023,39(18):201-210.
[11]贾璐,叶中华. 基于注意力机制和特征融合的葡萄病害识别模型[J]. 农业机械学报,2023,54(7):223-233.
[12]张林鍹,巴音塔娜,曾庆松. 基于StyleGAN2-ADA和改进YOLO v7的葡萄叶片早期病害检测方法[J]. 农业机械学报,2024,55(1):241-252.
[13]Verma S,Chug A,Singh A P,et al. PDS-MCNet:a hybrid framework using MobileNet v2 with SiLU6 activation function and capsule networks for disease severity estimation in plants[J]. Neural Computing and Applications,2023,35(25):18641-18664.
[14]Guo Y F,Lan Y T,Chen X D.CST:Convolutional Swin Transformer for detecting the degree and types of plant diseases[J]. Computers and Electronics in Agriculture,2022,202:107407.
[15]Liu X Y,Peng H W,Zheng N X,et al. EfficientViT:memory efficient vision transformer with cascaded group attention[C]//2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Vancouver,BC,Canada.2023:14420-14430.
[16]Shi D. TransNeXt:robust foveal visual perception for vision transformers[C]//2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Seattle,WA,USA.IEEE,2024:17773-17783.
[17]袁媛,陈雷. IDADP-葡萄病害识别研究图像数据集[J]. 中国科学数据,2022,7(1):86-90.
[18]Wu K,Peng H W,Chen M H,et al. Rethinking and improving relative position encoding for vision transformer[C]//2021 IEEE/CVF International Conference on Computer Vision (ICCV).Montreal,QC,Canada.IEEE,2021:10013-10021.
[19]Chen J R,Kao S H,He H,et al. Run,dont walk:chasing higher FLOPS for faster neural networks[C]//2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Vancouver,BC,Canada.IEEE,2023:12021-12031.
[20]Amirul Islam M,Kowal M,Jia S,et al. Global pooling,more than meets the eye:position information is encoded channel-wise in CNNs[C]//2021 IEEE/CVF International Conference on Computer Vision (ICCV).Montreal,QC,Canada.IEEE,2021:773-781.
[21]Dauphin Y N,Fan A,Auli M,et al. Language modeling with gated convolutional networks[C]//Proceedings of the 34th International Conference on Machine Learning,PMLR 70.Sydney,NSW,Australia.ACM,2017:933-941.
[22]Nayar S K,Narasimhan S G. Vision in bad weather[C]//Proceedings of the Seventh IEEE International Conference on Computer Vision..Kerkyra,Greece.IEEE,1999:820.
[23]Krizhevsky A,Sutskever I,Hinton G E. ImageNet classification with deep convolutional neural networks[J]. Communications of the ACM,2017,60(6):84-90.
[24]He K M,Zhang X Y,Ren S Q,et al. Deep residual learning for image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).Las Vegas,NV,USA.IEEE,2016:770-778.
[25]Howard A,Sandler M,Chen B,et al. Searching for MobileNet v3[C]//2019 IEEE/CVF International Conference on Computer Vision (ICCV).Seoul,Korea (South).IEEE,2019:1314-1324.
[26]Tan M X,Le Q V. EfficientNet:rethinking model scaling for convolutional neural networks[C]//Proceedings of the 36 th International Conference on MachineLearning,PMLR 97.Long Beach,California,2019.
[27]Zhou D Q,Kang B Y,Jin X J,et al. DeepViT:towards deeper vision transformer[EB/OL]. (2021-04-19)[2024-05-29].https://arxiv.org/abs/2103.11886v4.
[28]Graham B,El-Nouby A,Touvron H,et al. LeViT:a vision transformer in ConvNets clothing for faster inference[C]//2021 IEEE/CVF International Conference on Computer Vision (ICCV).Montreal,QC,Canada.IEEE,2021:12259-12269.
[29]Dosovitskiy A,Beyer L,Kolesnikov A,et al. An image is worth 16×16 words:Transformers for image recognition at scale[J/OL].ICLR,2021:1-21(2021-06-03)[2024-05-29].https://arxiv. org/abs/2010.11929.
[30]Liu Z,Lin Y T,Cao Y,et al. Swin transformer:hierarchical vision transformer using shifted windows[C]//2021 IEEE/CVF International Conference on Computer Vision (ICCV).Montreal,QC,Canada.IEEE,2021:9992-10002.
[31]Chen Y P,Dai X Y,Chen D D,et al. Mobile-former:Bridging mobilenet and transformer[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.2022:1-15.

相似文献/References:

[1]王利伟,徐晓辉,苏彦莽,等.基于计算机视觉的葡萄叶部病害识别研究[J].江苏农业科学,2017,45(23):222.
 Wang Liwei,et al.Study on recognition of grape leaf diseases based on computer vision[J].Jiangsu Agricultural Sciences,2017,45(5):222.
[2]戴久竣,马肄恒,吴坚,等.基于改进残差网络的葡萄叶片病害识别[J].江苏农业科学,2023,51(5):208.
 Dai Jiujun,et al.Grape leaf disease identification based on improved residual network[J].Jiangsu Agricultural Sciences,2023,51(5):208.
[3]许文燕.基于轻量化神经网络的葡萄叶部病害检测装置研制[J].江苏农业科学,2023,51(19):181.
 Xu Wenyan.Development of grape leaf disease detection device based on lightweight neural network[J].Jiangsu Agricultural Sciences,2023,51(5):181.
[4]张立强,武玲梅,蒋林利,等.基于改进YOLO v8s的葡萄叶片病害检测[J].江苏农业科学,2024,52(21):221.
 Zhang Liqiang,et al.Detection of grape leaf disease based on improved YOLO v8s[J].Jiangsu Agricultural Sciences,2024,52(5):221.

备注/Memo

备注/Memo:
收稿日期:2024-05-29
基金项目:山西省基础研究计划(编号:20210302123026)。
作者简介:刘宇雄(1998—),男,湖南长沙人,硕士,主要研究方向为智慧农业。E-mail:liuyuxiongzz@163.com。
通信作者:兰艳亭,博士,副教授,主要研究方向为智慧农业。E-mail:lytcyb@foxmail.com。
更新日期/Last Update: 2025-03-05