English

智能化农业装备学报(中英文) ›› 2025, Vol. 6 ›› Issue (2): 79-87.DOI: 10.12398/j.issn.2096-7217.2025.02.007

• • 上一篇    下一篇

基于改进VGG-16的烟叶成熟度识别

肖孟宇1(), 马云明1, 肖亦雄1, 陆峰2, 唐忠海3, 李跑3,4, 范伟3(), 肖航5   

  1. 1.湖南省烟草公司衡阳市公司,湖南 衡阳,421000
    2.湖北中烟工业有限责任公司,湖北 武汉,430040
    3.湖南农业大学食品科学技术学院,湖南 长沙,410128
    4.食品科学与生物技术湖南省重点实验室,湖南 长沙,410128
    5.麻省大学阿莫斯特分校食品科学系,美国马萨诸塞州阿莫斯特,01003
  • 收稿日期:2024-07-11 修回日期:2024-09-08 出版日期:2025-05-15 发布日期:2025-05-20
  • 通讯作者: 范伟
  • 作者简介:肖孟宇,男,1989年生,湖南邵阳人,助理农艺师;研究方向为烟草栽培。E-mail: 253875445@qq.com
  • 基金资助:
    国家自然科学基金项目(32360579);湖南省教育厅重点项目(20A230);湖南省烟草公司衡阳市公司科技项目(2021430481240017)

Fresh tobacco maturity recognition based on the improved VGG-16 model

XIAO Mengyu1(), MA Yunming1, XIAO Yixiong1, LU Feng2, TANG Zhonghai3, LI Pao3,4, FAN Wei3(), XIAO Hang5   

  1. 1.Hengyang Branch of Hunan Provincial Tobacco Company,Hengyang 421000,China
    2.Hubei Tobacco Industry Co. ,Ltd. ,Wuhan 430040,China
    3.College of Food Science and Technology,Hunan Agricultural University,Changsha 410128,China
    4.Hunan Province Key Laboratory of Food Science and Biotechnology,Changsha 410128,China
    5.Department of Food Science,University of Massachusetts-Amherst,Amherst,Massachusetts 01003,USA
  • Received:2024-07-11 Revised:2024-09-08 Online:2025-05-15 Published:2025-05-20
  • Contact: FAN Wei
  • Supported by:
    National Natural Science Foundation of China(32360579);Key Scientific Research Project of Hunan Provincial Department of Education(20A230);Science and Technology Project of Hunan Tobacco Company Hengyang Branch(2021430481240017);First Author:XIAO Mengyu, E-mail: 253875445@qq.com* Correspondence Author:FAN Wei, E-mail: weifan@hunau.edu.cn

摘要:

为提高采收时不同部位烟叶成熟度识别准确率,本研究系统收集上、中、下部3个部位烟叶的欠熟、适熟、过熟图像,提出一种基于深度学习的改进VGG-16的烟叶成熟度识别方法。该方法使用迁移学习预训练VGG-16模型作为基础模型,并对其进行微调、卷积层特征融合,以提升模型在烟叶成熟度识别任务上的性能,冻结卷积层及后续层的微调、添加BN层和使用Adam优化器,进一步提高了训练效率、避免过拟合现象产生并增强模型的稳健性和精确度。试验结果表明,改进后的VGG-16模型在烟叶成熟度识别任务中具有高准确度的优势,其测试集识别准确率达到99.7%,优于经典机器学习BP神经网络、支持向量机、原始VGG-16和VGG-19、AlexNet、ResNet50等方法结果。其参数量为14 721 353、模型大小为58.9 M、单张图像识别时间为0.024 9 s,表现出计算和存储资源需求低,识别速度快的高效优势。进一步通过Score-CAM算法对不同成熟度烟叶图像识别结果进行可视化分析,揭示了烟叶主脉中部区域为不同成熟度烟叶核心差异区域,为识别模型提供关键信息,进而可为明确不同部位烟叶成熟过程中化学物质转化规律提供信息。本研究提出的改进VGG-16深度学习模型在不同部位烟叶成熟度识别上具有高准确率和高效率的优势,有望为烟草采收生产提供精准有效的决策支持。未来可进一步探讨不同的特征融合策略和网络结构,以提高不同产地及不同年份间烟叶成熟度识别的泛化能力和稳健性。

关键词: 烟叶成熟度, 深度学习, 改进VGG-16, 迁移学习, 图像分类

Abstract:

To improve the accuracy of identifying the maturity of tobacco leaves at different harvesting stages, this study systematically collected images of under-ripe, ripe, and over-ripe tobacco leaves from the upper, middle, and lower parts of the plant. We proposed an improved VGG-16-based deep learning method for tobacco leaf maturity recognition. This method leverages a pre-trained VGG-16 model as its foundation, incorporating transfer learning, fine-tuning, and convolutional layer feature fusion to enhance the model's performance in tobacco leaf maturity recognition tasks. Freezing the convolutional and subsequent layers, adding BN layers, and using the Adam optimizer further improved training efficiency, avoided over fitting, and enhanced the model's robustness and accuracy. Experimental results showed that the improved VGG-16 model had a high accuracy advantage in tobacco leaf maturity recognition tasks, with a test set accuracy of 99.7%, surpassing classical machine learning methods such as BP neural networks, support vector machines, original VGG-16 and VGG-19, AlexNet, and ResNet50. The model comprised 14 721 353 parameters, a model size of 58.9 M, and a single image recognition time of 0.024 9 seconds, demonstrating the advantages of low computational and storage resource requirements and rapid recognition speed. Further visual analysis of the recognition results of tobacco leaf images with different maturities using the Score-CAM algorithm revealed that the central region of the main vein of tobacco leaves served as a primary differentiating feature for tobacco leaves with different maturities across different stalk positions, providing key information for the recognition model, there by clarifying the chemical substance transformation patterns occurring during the maturity process of tobacco leaves from distinct plant sections. The improved VGG-16 deep learning model proposed in this study exhibits high accuracy and efficiency in identifying the maturity of tobacco leaves at different harvesting stages. It is poised to provide precise and effective decision support for tobacco harvesting and production. Future research will focus on exploring alternative feature fusion strategies and network structures to further improve the generalization ability and robustness of tobacco leaf maturity recognition across different production areas and years.

Key words: tobacco maturity, deep learning, improved VGG-16, transfer learning, image classification

中图分类号: