基于微调卷积神经网络迁移学习模式下被害棉叶图像的识别

doi:10.6048/j.issn.1001-4330.2019.07.013

新疆农业科学 ›› 2019, Vol. 56 ›› Issue (7): 1294-1302.DOI: 10.6048/j.issn.1001-4330.2019.07.013

• 土壤肥料·贮藏加工·植物保护·畜牧水产 • 上一篇下一篇

基于微调卷积神经网络迁移学习模式下被害棉叶图像的识别

雷声渊¹, 马本学^1,2, 王文霞¹, 罗秀芝¹, 李玉洁¹, 戴建国³

1.石河子大学机械电气工程学院, 新疆石河子 832003;
2.农业部西北农业装备重点实验室,新疆石河子 832003;
3.石河子大学信息科学与技术学院,新疆石河子 832003

收稿日期:2019-04-22 出版日期:2019-07-20 发布日期:2019-09-05
通信作者: 马本学(1970-),男,山东人,教授,博士生导师,研究方向为图像光谱信息采集与处理技术、智能检测技术,(E-mail)mbx_shz@163.com
作者简介:雷声渊(1994-),男,陕西榆林人,硕士研究生,研究方向为农业信息化技术,(E-mail)lsy_shz@163.com
基金资助:
国家自然科学基金(31460317)

Preliminary Study on Image Recognition of Damaged Cotton Leaf Based on Fine-tuning Convolution Neural Network Transfer Learning Model

LEI Sheng-yuan¹, MA Ben-xue^1,2, WANG Wen-xia¹ LUO Xiu-zhi¹, LI Yu-jie¹, DAI Jian-guo³

1.College of Mechanical and Electrical Engineering, Shihezi University, Shihezi Xinjiang 832003,China;
2. Key Laboratory of Northwest Agricultural Equipment,Ministry of Agriculture, Shihezi Xinjiang 832003, China;
3. College of Information Science and Technology, Shihezi University, Shihezi Xinjiang 832003, China

Received:2019-04-22 Online:2019-07-20 Published:2019-09-05
Correspondence author: MA Ben-xue(1970-),male,native place:Sandong,china,professor,resear ch field.Image processing technology and intelligent detection technology, (E-mail)mbx_shz@163.com.
Supported by:
Supported by the National Natural Science Foundation of China "Research on the Temporal and Spatial Dynamic Monitoring Model of Cotton Leafhopper in Xinjiang Based on Multiple Environmental Factors"(31460317)

摘要/Abstract

摘要： 【目的】研究一种基于卷积神经网络的危害棉叶症状识别技术,提高棉花病虫害的识别准确率。【方法】基于caffe深度学习框架,在CaffeNet网络结构基础上增加一层全连接层(记为CaffeNet+1),并结合迁移学习方法对网络进行训练。采集健康、红叶茎枯、红蜘蛛、枯萎、黄萎、双斑萤叶甲、蚜虫、褐斑棉叶图像各975张作为样本集。随机选取验本集中80%的图像样本作为训练集,剩余20%作为测试集。【结果】迁移学习方式下学习率取0.005时的CaffeNet+1模型最优,在测试集上其识别准确率可达98.9%。【结论】在与全新学习模式下的CaffeNet模型相比,该方法可加速网络模型收敛,且具有更高的识别准确率,该技术方法在准确识别田间病虫害棉叶后表现症状的图像写出来具体方面具有重要的应用价值。

关键词: 卷积神经网络; 被害棉花; 病虫危害; 迁移学习; 图像识别

Abstract: 【Objective】 In order to improve the recognition accuracy of cotton diseases and insect pests, a method based on convolution neural network was proposed to identify the symptoms of harmful cotton leaves. The occurrence of cotton diseases and insect pests can cause severe yield loss in cotton. Identification of diseases and pests is key to the prevention and control of diseases and pest. 【Method】To improve the identification accuracy of cotton diseases and insect pests, a method for identification cotton leaf disease and pest based on improve convolutional neural network and transfer learning were proposed. 【Result】The experimental results showed that in the dataset containing 7,800 images of healthy cotton leaves and leaves of seven kinds of diseases and pests (divided into training set and test set at the ratio of 4∶1), under the transfer learning method, the CaffeNet+1 model with the learning rate of 0.005 was optimal, and the recognition accuracy of the test set was 98.9%. 【Conclusion】Compared with the CaffeNet model under the new learning model, this method can accelerate the convergence of the network model and has higher recognition accuracy. This technique has important application value in accurately identifying the image of symptoms of pests and diseases of cotton leaves in the field.

Key words: convolutional neural network; damaged cotton; disease and pest harm; transfer learning; image recognition

中图分类号:

雷声渊, 马本学, 王文霞, 罗秀芝, 李玉洁, 戴建国. 基于微调卷积神经网络迁移学习模式下被害棉叶图像的识别[J]. 新疆农业科学, 2019, 56(7): 1294-1302.

LEI Sheng-yuan, MA Ben-xue, WANG Wen-xia LUO Xiu-zhi, LI Yu-jie, DAI Jian-guo. Preliminary Study on Image Recognition of Damaged Cotton Leaf Based on Fine-tuning Convolution Neural Network Transfer Learning Model[J]. Xinjiang Agricultural Sciences, 2019, 56(7): 1294-1302.

参考文献 22

[1]	Khana M, Damalasb CA. (2015). Farmers' knowledge about common pests and pesticide safety in conventional cotton production in Pakistan . Crop Protection, 77: 45 - 51.
[2]	朱荷琴.棉花主要病害研究概要[J].棉花学报,2007,(5):391-398.ZHU He-qin. (2007).A Summary of Researches on Main Cotton Diseases [J]. Cotton Science, (5): 391 - 398.(in Chinese)
[3]	王献锋,张传雷,张善文,等. 基于自适应判别深度置信网络的棉花病虫害预测[J].农业工程学报,2018,34(14):157-164.WANG Xian-feng, ZHANG Chuan-lei, ZHANG Shan-wen, et al. (2018). Forecasting of cotton diseases and pests based on adaptive discriminant deep belief network [J]. Transactions of the CSAE, 34(14): 157 - 164. (in Chinese)
[4]	张建华,冀荣华,袁雪,等.基于径向基支持向量机的棉花虫害识别[J]. 农业机械学报, 2011, 42(8) :178 - 183.ZHANG Jian-hua, JI Rong-hua, YUAN Xue, et al. (2011).Recognition of Pest Damage for Cotton Leaf Based on RBF － SVM Algorithm[J]. Transactions of the Chinese Society for Agricultural Machinery, 48(3): 30 - 37.(in Chinese)
[5]	Bernardes A A, Rogeri J G, Marranghello N, et al. (2013).Identification of foliar diseases in cotton crop . Topics in Medical Image Processing & Computational Vision:67 - 85.
[6]	翟治芬,徐哲,周新群,等. 基于朴素贝叶斯分类器的棉花盲椿象危害等级识别[J]. 农业工程学报, 2015, 31(1) : 204 - 211.ZHAI Zhi-fen, XU Zhe, ZHOU Xin-qun, et al. (2015).Recognition of hazard grade for cotton blind stinkbug based on Naive Bayesian classifier [J]. Transactions of the CSAE, 31(1) : 204 - 211.(in Chinese)
[7]	Hinton G E, Salakhutdinov R R. (2006).Reducing the dimensionality of data with neural network . Science, 313(5786): 50 - 507.
[8]	Guan W, Yu S, Jianxin W. (2017).Automatic Image-Based Plant Disease Severity Estimation Using Deep Learning . Computational Intelligence and Neuroscience, 2017:1-8.
[9]	杨国国,鲍一丹,刘子毅. 基于图像显著性分析与卷积神经网络的茶园害虫定位与识别[J]. 农业工程学报, 33(6) : 156 - 162.YANG Guo-guo, BAO Yi-dan, LIU Zi-yi. (2017).Localization and recognition of pests in tea plantation based on image saliency analysis and convolutional neural network[J] . Transactions of the CSAE, 33(6):156 - 162.(in Chinese)
[10]	Lu Y, Yi S J, Zeng N Y, et al. (2017).Identification of rice diseases using deep convolutional neural networks . Neurocomputing, 267 : 378 - 384.
[11]	Liu B, Zhang Y, He D J, et al. (2017).Identification of Apple Leaf Diseases Based on Deep Convolutional Neural Networks . SYMMETRY-BASEL, 10(1) : 11.
[12]	谭文学, 赵春江, 吴华瑞, 等. 基于弹性动量深度学习神经网络的果体病理图像识别[J]. 农业机械学报, 2015, 46(1) : 20 - 25.TAN Wen-xue,ZHAO Chun-jiang,WU Hua-rui,et al. (2015).A Deep Learning Network for Recognizing Fruit Pathologic Images Based on Flexible Momentum [J]. Transactions of the Chinese Society for Agricultural Machinery, 46(1) : 20 - 25. (in Chinese)
[13]	赵冰梅,李贤超,王俊刚. 2011年新疆兵团棉花病虫害发生特点及原因分析[J]. 中国棉花, 2012, 39(3) : 9 - 11.ZHAO Bing-mei, LI Xian-chao, WANG Jun-gang. (2012). Analysis of the characteristics and causes of cotton diseases and insect pests in middle and later stage of Xinjiang production and construction crops[J] . Chinese Cotton, 39(3) : 9 - 11.(in Chinese)
[14]	雒珺瑜,马艳,崔金杰,等. 棉花病虫害诊断及防治原色图谱[M]. 北京:金盾出版社,2015 : 15 - 97.LUO Jun-yu, MA Yang, CUI Jin-jie, et al. (2015)Diagnosis and Control Primitive Color Map of Cotton Diseases and Pests [M]. Beijing: Golden Shield Publish House, 15 - 97. (in Chinese)
[15]	Jia Y, Shelhamer E, Donahue J , et al. (2014 ).Caffe: Convolutional Architecture for Fast Feature Embedding . Acm International Conference on Multimedia, 675 - 678.
[16]	Krizhevsky A , Sutskever I , Hinton G . (2017). ImageNet Classification with Deep Convolutional Neural Networks NIPS. Curran Associates Inc. 2012. Krizhevsky A, Sutskever I, Hinton G E. ImageNet Classification with deep convolutional neural networks . Communications of the ACM, 60(6) : 84 - 90.
[17]	Hagan M T, Demuth H B, Beale M H.戴葵,译. 神经网络设计 [M]. 北京:机械工业出版社,2002.Hagan M T, Demuth H B, Beale M H. (2002). Neural Network Design Translated by Dai Kui [M]. Beijing: China Machine Press, (in Chinese)
[18]	Hinton G E. (2010).Rectified Linear Units Improve Restricted Boltzmann Machines Vinod Nair International Conference on International Conference on Machine Learning. Omnipress, 807 - 814.
[19]	Lecun Y , Bengio Y , Hinton G . (2015).Deep learning.[J]. Nature, 521(7553):436. - 444.
[20]	Scherer D, Müller A, Behnke S. (2010). Evaluation of Pooling Operations in Convolutional Architectures for Object Recognition . Artificial Neural Networks - ICANN, 6,354 : 92 - 101.
[21]	Guan W , Yu S , Jianxin W . (2017). Automatic Image-Based Plant Disease Severity Estimation Using Deep Learning [J]. Computational Intelligence and Neuroscience: 1-8.
[22]	Behbahan Navid Samimi, Mousavinasab Zohreh. (2012).Presenting a method based on back-propagation neural network for recognition of handwritten digits . International Journal of Management, IT and Engineering, 2(12):3-73.

基于微调卷积神经网络迁移学习模式下被害棉叶图像的识别

Preliminary Study on Image Recognition of Damaged Cotton Leaf Based on Fine-tuning Convolution Neural Network Transfer Learning Model

PDF下载

可视化

摘要/Abstract

引用本文

使用本文

参考文献 22

相关文章 2

编辑推荐

Metrics

[1]	李雨晴, 陈燕红, 李永可, 肖天赐, 李清源. 作物害虫图像智能识别方法[J]. 新疆农业科学, 2023, 60(12): 2973-2981.
[2]	罗秀芝, 马本学, 李小霞, 胡洋洋, 王文霞, 雷声渊. 基于卷积神经网络干制哈密大枣纹理分级[J]. 新疆农业科学, 2018, 55(12): 2220-2227.