PMVT: a lightweight vision transformer for plant disease identification on mobile devices
文献类型: 外文期刊
作者: Li, Guoqiang 1 ; Wang, Yuchao 2 ; Zhao, Qing 1 ; Yuan, Peiyan 2 ; Chang, Baofang 2 ;
作者机构: 1.Henan Acad Agr Sci, Inst Agr Econ & Informat, Zhengzhou, Henan, Peoples R China
2.Henan Normal Univ, Coll Comp & Informat Engn, Xinxiang, Henan, Peoples R China
3.Key Lab Artificial Intelligence & Personalized Lea, Xinxiang, Henan, Peoples R China
关键词: plant disease identification; vision transformer; lightweight model; attention module; APP
期刊名称:FRONTIERS IN PLANT SCIENCE ( 影响因子:5.6; 五年影响因子:6.8 )
ISSN: 1664-462X
年卷期: 2023 年 14 卷
页码:
收录情况: SCI
摘要: Due to the constraints of agricultural computing resources and the diversity of plant diseases, it is challenging to achieve the desired accuracy rate while keeping the network lightweight. In this paper, we proposed a computationally efficient deep learning architecture based on the mobile vision transformer (MobileViT) for real-time detection of plant diseases, which we called plant-based MobileViT (PMVT). Our proposed model was designed to be highly accurate and low-cost, making it suitable for deployment on mobile devices with limited resources. Specifically, we replaced the convolution block in MobileViT with an inverted residual structure that employs a 7x7 convolution kernel to effectively model long-distance dependencies between different leaves in plant disease images. Furthermore, inspired by the concept of multi-level attention in computer vision tasks, we integrated a convolutional block attention module (CBAM) into the standard ViT encoder. This integration allows the network to effectively avoid irrelevant information and focus on essential features. The PMVT network achieves reduced parameter counts compared to alternative networks on various mobile devices while maintaining high accuracy across different vision tasks. Extensive experiments on multiple agricultural datasets, including wheat, coffee, and rice, demonstrate that the proposed method outperforms the current best lightweight and heavyweight models. On the wheat dataset, PMVT achieves the highest accuracy of 93.6% using approximately 0.98 million (M) parameters. This accuracy is 1.6% higher than that of MobileNetV3. Under the same parameters, PMVT achieved an accuracy of 85.4% on the coffee dataset, surpassing SqueezeNet by 2.3%. Furthermore, out method achieved an accuracy of 93.1% on the rice dataset, surpassing MobileNetV3 by 3.4%. Additionally, we developed a plant disease diagnosis app and successfully used the trained PMVT model to identify plant disease in different scenarios.
- 相关文献
作者其他论文 更多>>
-
A general-purpose edge-feature guidance module to enhance vision transformers for plant disease identification
作者:Chang, Baofang;Wang, Yuchao;Zhao, Xiaoyan;Yuan, Peiyan;Chang, Baofang;Wang, Yuchao;Zhao, Xiaoyan;Yuan, Peiyan;Chang, Baofang;Wang, Yuchao;Zhao, Xiaoyan;Yuan, Peiyan;Li, Guoqiang
关键词:Plant disease identification; Vision transformers; Edge features; Deep learning
-
One-off irrigation improves water and nitrogen use efficiency and productivity of wheat as mediated by nitrogen rate and tillage in drought-prone areas
作者:Zhao, Kainan;Wang, Hongtao;Wu, Jinzhi;Huang, Xiuli;Wu, Shanwei;Zhang, Jun;Zhang, Zhenwang;Hou, Yuanquan;Zhao, Zhiming;Li, Shuang;Guo, Jinhua;Zhao, Wenxin;Li, Shujing;Li, Wenna;Huang, Ming;Li, Youjun;Liu, Akang;Li, Guoqiang
关键词:One-off irrigation; Sub-soiling; Nitrogen; Yield; Nitrate-N residue
-
Alternative Furrow Irrigation Combined with Topdressing Nitrogen at Jointing Help Yield Formation and Water Use of Winter Wheat under No-Till Ridge Furrow Planting System in Semi-Humid Drought-Prone Areas of China
作者:Wu, Jinzhi;Guan, Haoyue;Li, Youjun;Fu, Guozhan;Huang, Ming;Wang, Zhimin;Li, Guoqiang;Li, Guoqiang
关键词:semi-humid drought-prone areas; winter wheat; alternative furrow irrigation; topdressing N; grain yield; WUE
-
Detection method of wheat spike improved YOLOv5s based on the attention mechanism
作者:Zang, Hecang;Zhou, Meng;Chen, Dandan;Zhao, Qing;Zhang, Jie;Li, Guoqiang;Zheng, Guoqing;Zang, Hecang;Zhou, Meng;Chen, Dandan;Zhao, Qing;Zhang, Jie;Li, Guoqiang;Zheng, Guoqing;Wang, Yanjing;Ru, Linyuan
关键词:wheat; spike number detection; attention mechanism; deep learning; YOLOv5s
-
Estimation of Density and Height of Winter Wheat Varieties Using Unmanned Aerial Vehicles Images
作者:Zang, Hecang;Yang, Xiuzhong;He, Jia;Zhou, Meng;Zheng, Guoqing;Li, Guoqiang;Zang, Hecang;Yang, Xiuzhong;He, Jia;Zhou, Meng;Zheng, Guoqing;Li, Guoqiang;Zang, Hecang;Yang, Xiuzhong;He, Jia;Zhou, Meng;Zheng, Guoqing;Li, Guoqiang;Wang, Yanjing
关键词:Unmanned Aerial Vehicle; New Wheat Varieties; Coverage; Plant Density; Plant Height
-
The design and experiment of peanut high-throughput automatic seed testing system based on machine learning
作者:Zang, Hecang;Zhao, Qiaoli;Zhao, Qing;Zhang, Jie;Wang, Meng;Zheng, Guoqing;Li, Guoqiang;Zang, Hecang;Zhao, Qiaoli;Zhao, Qing;Zhang, Jie;Wang, Meng;Zheng, Guoqing;Li, Guoqiang;Zang, Hecang;Zhao, Qiaoli;Zhao, Qing;Zhang, Jie;Wang, Meng;Zheng, Guoqing;Li, Guoqiang;Wang, Yanjing
关键词:Peanut; machine learning; high-throughput; seed testing system; experiment
-
Field road segmentation network based on PraNet
作者:Liu, Guoqi;Zhao, Manqi;Bai, Lu;Chang, Baofang;Zang, Hecang
关键词:Road extraction; CNN; Segmentation