基于迁移学习与卷积神经网络的玉米植株病害识别
|
摘要
大数据背景下产生了海量图像数据,传统的图像识别方法识别玉米植株病害准确率较低,已远远不能满足需求。卷积神经网络作为深度学习中的常用算法被广泛用于处理机器视觉问题,能自动识别和提取图像特征。因此,本研究提出一种基于数据增强与迁移学习相结合的卷积神经网络识别玉米植株病害模型。该算法首先通过数据增强方法增加数据,以提高模型的泛化性和准确率;再构建基于迁移学习的卷积神经网络模型,引入该模型的训练方式,提取病害图片特征,加速卷积神经网络的训练过程,降低网络的过拟合程度;最后将该模型运用到从农田采集的玉米病害图片,进行玉米病害的精确识别。识别试验结果表明:使用数据增强与迁移学习的卷积神经网络优化算法对玉米主要病害(玉米大斑病、小斑病、灰斑病、黑穗病及瘤黑粉病)的平均识别准确度达96.6%,和单一的卷积神经网络相比,精度提高了25.6%,处理每张图片时间为0.28s,比传统神经网络缩短了将近10倍。本算法的精确度和训练速度上比传统卷积神经网络有明显提高,为玉米等农作物植株病害的识别提供了新方法。
Corn is one of the most important food crops in China,and the occurrence of disease will result in serious yield reduction.Therefore,the diagnosis and treatment of corn disease is an important link in corn production.Under the background of big data,massive image data are generated.The traditional image recognition method has a low accuracy in identifying corn plant diseases,which is far from meeting the needs.With the development of artificial intelligence and deep learning,convolutional neural network,as a common algorithm in deep learning,is widely used to deal with machine vision problems.It can automatically identify and extract image features.However,in image classification,CNN still has problems such as small sample size,high sample similarity and long training convergence time.CNN has the limitations of expression ability and lack of feedback mechanism,and data enhancement and transfer learning can solve the corresponding problems.Therefore,this research proposed an optimization algorithm for corn plant disease recognition based on the convolution neural network recognition model combining data enhancement and transfer learning.Firstly,the algorithm preprocessed the data through the data enhancement method to expand the data set,so as to improve the generalization and accuracy of the model.Then,the CNN model based on transfer learning was constructed.The Inception V3 model was adopted through transfer learning to extract the image characteristics of the disease while keeping the parameters unchanged.In this way,the training process of the convolutional neural network was accelerated and the over-fitting degree of the network was reduced.The extracted image features were used as input of the CNN to train the network,and finally the recognition results were obtained.Finally,the model was applied to the pictures of corn diseases collected from the farmland to accurately identify five kinds of corn diseases.Identification test results showed that using data to enhance the CNN optimization algorithm and the migration study on the average recognition accuracy main diseases of com(spot,southern leaf blight,gray leaf spot,smut,gall smut) reached 96.6%,which compared with single CNN,has greatly improved the precision and identification precision by 25.6% on average.The average processing time of each image was 0.28 s,shortens nearly 10 times than a single convolution neural network.The experimental results show that the algorithm is more accurate and faster than the traditional CNN,which provides a new method for identification of corn plant diseases.
Corn is one of the most important food crops in China,and the occurrence of disease will result in serious yield reduction.Therefore,the diagnosis and treatment of corn disease is an important link in corn production.Under the background of big data,massive image data are generated.The traditional image recognition method has a low accuracy in identifying corn plant diseases,which is far from meeting the needs.With the development of artificial intelligence and deep learning,convolutional neural network,as a common algorithm in deep learning,is widely used to deal with machine vision problems.It can automatically identify and extract image features.However,in image classification,CNN still has problems such as small sample size,high sample similarity and long training convergence time.CNN has the limitations of expression ability and lack of feedback mechanism,and data enhancement and transfer learning can solve the corresponding problems.Therefore,this research proposed an optimization algorithm for corn plant disease recognition based on the convolution neural network recognition model combining data enhancement and transfer learning.Firstly,the algorithm preprocessed the data through the data enhancement method to expand the data set,so as to improve the generalization and accuracy of the model.Then,the CNN model based on transfer learning was constructed.The Inception V3 model was adopted through transfer learning to extract the image characteristics of the disease while keeping the parameters unchanged.In this way,the training process of the convolutional neural network was accelerated and the over-fitting degree of the network was reduced.The extracted image features were used as input of the CNN to train the network,and finally the recognition results were obtained.Finally,the model was applied to the pictures of corn diseases collected from the farmland to accurately identify five kinds of corn diseases.Identification test results showed that using data to enhance the CNN optimization algorithm and the migration study on the average recognition accuracy main diseases of com(spot,southern leaf blight,gray leaf spot,smut,gall smut) reached 96.6%,which compared with single CNN,has greatly improved the precision and identification precision by 25.6% on average.The average processing time of each image was 0.28 s,shortens nearly 10 times than a single convolution neural network.The experimental results show that the algorithm is more accurate and faster than the traditional CNN,which provides a new method for identification of corn plant diseases.
引文
[1]王晓鸣.六种重要玉米病害病原名称的厘定[J].中国农业科学,2018,51(18):3497-3507.Wang X.Collation of scientific names of six maize disease pathogens.[J].Scientia Agricultura Sinica,2018,51(18):3497-3507.
[2]付立思,何荣卜,刘朋维.一种新的不变矩与神经网络玉米病害识别系统[J].计算机工程与应用,2012,48(04):219-221.Fu L,He R,Liu P.New system about moment invariant and neural network used in maize disease recognition[J].Computer Engineering and Applications,2012,48(04):219-221.
[3]谭峰,马晓丹.基于叶片的植物病虫害识别方法[J].农机化研究,2009,31(06):41-43.Tan F,Ma X.The method of recognition of damage by disease and insect based on laminae[J].Journal of Agricultural Mechanization Research,2009,31(06):41-43.
[4]田有文.基于纹理特征和支持向量机的葡萄病害的识别[J].仪器仪表学报,2005(S1):606-608.Tian Y.A method of recognizing grape disease based on texture feature and support vector machine[J].Chinese Journal of Scientific Instrument,2005(S1):606-608.
[5]王献锋,张善文,孔韦韦.基于自适应中心对称局部二值模式的作物病害识别方法[J].广东农业科学,2016,43(09):140-145.Wang X,Zhang S,Kong W.Adaptive cen ter-symmetric local binary patterns for crop disease recognition[J].Guangdong Agricultural Sciences,2016,43(09):140-145.
[6]Zhang S,Shang Y,Wang L.Plant disease recognition based on plant leaf image[J].Journal of Animal and Plant Sciences,2015,25(3):42-45.
[7]付立思,何荣卜,刘朋维.一种新的不变矩与神经网络玉米病害识别系统[J].计算机工程与应用,2012,48(04):219-221.Fu L,He R,Liu P.New system about moment invariant and neural network used in maize disease recognition[J].Computer Engineering and Applications,2012,48(04):219-221.
[8]马宗帅.基于深度学习的心脑血管疾病预测方法研究[D].西安:西安建筑科技大学,2015.Ma Z.Research on cardiovascular disease pre diction based on Deep Learning technical[D].Xi’an:Xi’an University of Architecture and Technology,2015.
[9]孙俊,谭文军,毛罕平.基于改进卷积神经网络的多种植物叶片病害识别[J].农业工程学报,2017,33(19):209-215.Sun J,Tan W,Mao H.Recognition of multiple plant leaf diseases based on improved convolutional neural network[J].Transactions of the CSAE,2017,33(19):209-215.
[10]李敬.基于卷积神经网络的烟草病害自动识别研究[D].泰安:山东农业大学,2016.Li J.Automatci identification research of tobacco disease based on convolutional neural network[D].Tai'an:Shandong Agricultural University,2016.
[11]邓立苗,唐俊,马文杰.基于图像处理的玉米叶片特征提取与识别系统[J].中国农机化学报,2014,35(06):72-75,79.Deng L,Tang J,Ma W.Feature extraction and recognition system of maize leaf based on image processing[J].Journal of Chinese Agricultural Mechanization,2014,35(06):72-75,79.
[12]王振,高茂庭.基于卷积神经网络的图像识别算法设计与实现[J].现代计算机(专业版),2015(20):61-66.Wang Z,Gao M.Design and implementation of image recognition algorithm based on convolutional neural networks[J].Modern Computer,2015(20):61-66.
[13]曹丽英,张晓贤,伞晓辉,等.基于图像处理技术和BP神经网络算法的玉米病害诊断方法的研究[J].计算机科学,2012,39(10):300-302.Cao L,Zhang X,San X,et al.Research on maize disease diagnosis method based on image processing technique and BP neural network algorithm[J].Computer Science,2012,39(10):300-302.
[14]罗建华,李明奇,郑泽忠,等.基于深度卷积神经网络的高光谱遥感图像分类[J].西华大学学报(自然科学版),2017,36(04):13-20.Luo J,Li M,Zheng Z,et al.Hyperspectral remote sensing images classification using a deep convolutional neural network model[J].Journal of Xihua University(Natural Science Edition),2017,36(04):13-20.
[15]石祥滨,房雪键,张德园,等.基于深度学习混合模型迁移学习的图像分类[J].系统仿真学报,2016,28(01):167-173,182.Shi X,Fang X,Zhang D,et al.Image classification based on mixed deep learning model transfer learning[J].Journal of System Simulation,2016,28(01):167-173,182.
[16]李彦冬,郝宗波,雷航.卷积神经网络研究综述[J].计算机应用,2016,36(09):2508-2515,2565.Li Y,Hao,Lei H.Survey of convolutional neu ral network[J].Journal of Computer Applications,2016,36(09):2508-2515,2565.
[17]石祥滨,房雪键,张德园,等.基于深度学习混合模型迁移学习的图像分类[J].系统仿真学报,2016,28(01):167-173,182.Shi X,Fang X,Zhang D,et al.Image classification based on mixed deep learning model transfer learning[J].Journal of System Simulation,2016,28(01):167-173,182.
[18]周飞燕,金林鹏,董军.卷积神经网络研究综述[J].计算机学报,2017,40(06):1229-1251.Zhou F,Jin L,Dong J.Review of convolutional neural network[J].Chinese Journal of Computers,2017,40(06):1229-1251.
[19]王振,高茂庭.基于卷积神经网络的图像识别算法设计与实现[J].现代计算机(专业版),2015(20):61-66.Wang Z,Gao M.Design and implementation of image recognition algorithm based on convolutional neural networks[J].Modern Computer,2015(20):61-66.
[20]蒋帅.基于卷积神经网络的图像识别[D].长春:吉林大学,2017.Jiang S.Image recongnition based on convolution neural networks[D].Changchun:Jilin University,2017.
[21]张帅,淮永建.基于分层卷积深度学习系统的植物叶片识别研究[J].北京林业大学报,2016,38(09):108-115.Zhang S,Huai Y.Leaf image recognition based on layered convolutions neural network deep learning[J].Journal of Beijing Forestry University,2016,38(09):108-115.
[22]王献锋,张善文,王震,等.基于叶片图像和环境信息的黄瓜病害识别方法[J].农业工程学报,2014,30(14):148-153.Wang X,Zhang S,Wang Z,et al.Recognition of cucumber diseases based on leaf image and environmental information[J].Transactions of the CSAE 2014,30(14):148-153.
[23]高友文,周本君,胡晓飞.基于数据增强的卷积神经网络图像识别研究[J].计算机技术与发展,2018(08):1-6.Gao Y,Zhou B,Hu X.Research on image recognition of convolution neural network based on data enhancement[J].Computer Technology and De velopment,2018(08):1-6.
[24]褚晶辉,吴泽蕤,吕卫,等.基于迁移学习和深度卷积神经网络的乳腺肿瘤诊断系统[J].激光与光电子学进展,2018,55(8):081001,1-12.Chu J,Wu Z,Lv W,et al.Breast cancer diagnosis system based on transfer learning and deep convolutional neural networks[J].Laser&Optoelectronics Progress,2018,55(8):081001,1-12.
[25]米硕,孙瑞彬,李欣,等.迁移学习在卷积神经网络中的应用[J].中国新通信,2018,20(07):121.Mi S,Sun R,Li X,et al.The application of transfer learning in convolution neural network[J].China New Telecommunications,2018,20(07):121.
[26]鲁恒,付萧,贺一楠,等.基于迁移学习的无人机影像耕地信息提取方法[J].农业机械学报,2015,46(12):274-279,284.Lu H,Fu X,He Y,et al.Cultivated land information extraction from high resolution UAV images based on transfer learning[J].Transactions of the CSAM,2015,46(12):274-279,284.
[27]Oquab M,Bottou L,Laptev I.Learning and transferring mid-level image representation using convolutional neural network[C].The 27th IEEE conference on Computer Vision and Pattern Recognition,Columbus,USA,2014.
[28]林宇,陈晓勇.基于inception v3模型的道路交通标志识别研究[J].江西科学,2018,36(05):849-852.Lin Y,Chen X.Inception-v3 for road traffic signs recognition[J].Jiangxi Science,2018,36(05):849-852.
[29]宋光慧.基于迁移学习与深度卷积特征的图像标注方法研究[D].杭州:浙江大学,2017.Song G.Image annotation method based on transfer learning and deep convolutional feature[D].Hangzhou:Zhejiang University,2017.
[2]付立思,何荣卜,刘朋维.一种新的不变矩与神经网络玉米病害识别系统[J].计算机工程与应用,2012,48(04):219-221.Fu L,He R,Liu P.New system about moment invariant and neural network used in maize disease recognition[J].Computer Engineering and Applications,2012,48(04):219-221.
[3]谭峰,马晓丹.基于叶片的植物病虫害识别方法[J].农机化研究,2009,31(06):41-43.Tan F,Ma X.The method of recognition of damage by disease and insect based on laminae[J].Journal of Agricultural Mechanization Research,2009,31(06):41-43.
[4]田有文.基于纹理特征和支持向量机的葡萄病害的识别[J].仪器仪表学报,2005(S1):606-608.Tian Y.A method of recognizing grape disease based on texture feature and support vector machine[J].Chinese Journal of Scientific Instrument,2005(S1):606-608.
[5]王献锋,张善文,孔韦韦.基于自适应中心对称局部二值模式的作物病害识别方法[J].广东农业科学,2016,43(09):140-145.Wang X,Zhang S,Kong W.Adaptive cen ter-symmetric local binary patterns for crop disease recognition[J].Guangdong Agricultural Sciences,2016,43(09):140-145.
[6]Zhang S,Shang Y,Wang L.Plant disease recognition based on plant leaf image[J].Journal of Animal and Plant Sciences,2015,25(3):42-45.
[7]付立思,何荣卜,刘朋维.一种新的不变矩与神经网络玉米病害识别系统[J].计算机工程与应用,2012,48(04):219-221.Fu L,He R,Liu P.New system about moment invariant and neural network used in maize disease recognition[J].Computer Engineering and Applications,2012,48(04):219-221.
[8]马宗帅.基于深度学习的心脑血管疾病预测方法研究[D].西安:西安建筑科技大学,2015.Ma Z.Research on cardiovascular disease pre diction based on Deep Learning technical[D].Xi’an:Xi’an University of Architecture and Technology,2015.
[9]孙俊,谭文军,毛罕平.基于改进卷积神经网络的多种植物叶片病害识别[J].农业工程学报,2017,33(19):209-215.Sun J,Tan W,Mao H.Recognition of multiple plant leaf diseases based on improved convolutional neural network[J].Transactions of the CSAE,2017,33(19):209-215.
[10]李敬.基于卷积神经网络的烟草病害自动识别研究[D].泰安:山东农业大学,2016.Li J.Automatci identification research of tobacco disease based on convolutional neural network[D].Tai'an:Shandong Agricultural University,2016.
[11]邓立苗,唐俊,马文杰.基于图像处理的玉米叶片特征提取与识别系统[J].中国农机化学报,2014,35(06):72-75,79.Deng L,Tang J,Ma W.Feature extraction and recognition system of maize leaf based on image processing[J].Journal of Chinese Agricultural Mechanization,2014,35(06):72-75,79.
[12]王振,高茂庭.基于卷积神经网络的图像识别算法设计与实现[J].现代计算机(专业版),2015(20):61-66.Wang Z,Gao M.Design and implementation of image recognition algorithm based on convolutional neural networks[J].Modern Computer,2015(20):61-66.
[13]曹丽英,张晓贤,伞晓辉,等.基于图像处理技术和BP神经网络算法的玉米病害诊断方法的研究[J].计算机科学,2012,39(10):300-302.Cao L,Zhang X,San X,et al.Research on maize disease diagnosis method based on image processing technique and BP neural network algorithm[J].Computer Science,2012,39(10):300-302.
[14]罗建华,李明奇,郑泽忠,等.基于深度卷积神经网络的高光谱遥感图像分类[J].西华大学学报(自然科学版),2017,36(04):13-20.Luo J,Li M,Zheng Z,et al.Hyperspectral remote sensing images classification using a deep convolutional neural network model[J].Journal of Xihua University(Natural Science Edition),2017,36(04):13-20.
[15]石祥滨,房雪键,张德园,等.基于深度学习混合模型迁移学习的图像分类[J].系统仿真学报,2016,28(01):167-173,182.Shi X,Fang X,Zhang D,et al.Image classification based on mixed deep learning model transfer learning[J].Journal of System Simulation,2016,28(01):167-173,182.
[16]李彦冬,郝宗波,雷航.卷积神经网络研究综述[J].计算机应用,2016,36(09):2508-2515,2565.Li Y,Hao,Lei H.Survey of convolutional neu ral network[J].Journal of Computer Applications,2016,36(09):2508-2515,2565.
[17]石祥滨,房雪键,张德园,等.基于深度学习混合模型迁移学习的图像分类[J].系统仿真学报,2016,28(01):167-173,182.Shi X,Fang X,Zhang D,et al.Image classification based on mixed deep learning model transfer learning[J].Journal of System Simulation,2016,28(01):167-173,182.
[18]周飞燕,金林鹏,董军.卷积神经网络研究综述[J].计算机学报,2017,40(06):1229-1251.Zhou F,Jin L,Dong J.Review of convolutional neural network[J].Chinese Journal of Computers,2017,40(06):1229-1251.
[19]王振,高茂庭.基于卷积神经网络的图像识别算法设计与实现[J].现代计算机(专业版),2015(20):61-66.Wang Z,Gao M.Design and implementation of image recognition algorithm based on convolutional neural networks[J].Modern Computer,2015(20):61-66.
[20]蒋帅.基于卷积神经网络的图像识别[D].长春:吉林大学,2017.Jiang S.Image recongnition based on convolution neural networks[D].Changchun:Jilin University,2017.
[21]张帅,淮永建.基于分层卷积深度学习系统的植物叶片识别研究[J].北京林业大学报,2016,38(09):108-115.Zhang S,Huai Y.Leaf image recognition based on layered convolutions neural network deep learning[J].Journal of Beijing Forestry University,2016,38(09):108-115.
[22]王献锋,张善文,王震,等.基于叶片图像和环境信息的黄瓜病害识别方法[J].农业工程学报,2014,30(14):148-153.Wang X,Zhang S,Wang Z,et al.Recognition of cucumber diseases based on leaf image and environmental information[J].Transactions of the CSAE 2014,30(14):148-153.
[23]高友文,周本君,胡晓飞.基于数据增强的卷积神经网络图像识别研究[J].计算机技术与发展,2018(08):1-6.Gao Y,Zhou B,Hu X.Research on image recognition of convolution neural network based on data enhancement[J].Computer Technology and De velopment,2018(08):1-6.
[24]褚晶辉,吴泽蕤,吕卫,等.基于迁移学习和深度卷积神经网络的乳腺肿瘤诊断系统[J].激光与光电子学进展,2018,55(8):081001,1-12.Chu J,Wu Z,Lv W,et al.Breast cancer diagnosis system based on transfer learning and deep convolutional neural networks[J].Laser&Optoelectronics Progress,2018,55(8):081001,1-12.
[25]米硕,孙瑞彬,李欣,等.迁移学习在卷积神经网络中的应用[J].中国新通信,2018,20(07):121.Mi S,Sun R,Li X,et al.The application of transfer learning in convolution neural network[J].China New Telecommunications,2018,20(07):121.
[26]鲁恒,付萧,贺一楠,等.基于迁移学习的无人机影像耕地信息提取方法[J].农业机械学报,2015,46(12):274-279,284.Lu H,Fu X,He Y,et al.Cultivated land information extraction from high resolution UAV images based on transfer learning[J].Transactions of the CSAM,2015,46(12):274-279,284.
[27]Oquab M,Bottou L,Laptev I.Learning and transferring mid-level image representation using convolutional neural network[C].The 27th IEEE conference on Computer Vision and Pattern Recognition,Columbus,USA,2014.
[28]林宇,陈晓勇.基于inception v3模型的道路交通标志识别研究[J].江西科学,2018,36(05):849-852.Lin Y,Chen X.Inception-v3 for road traffic signs recognition[J].Jiangxi Science,2018,36(05):849-852.
[29]宋光慧.基于迁移学习与深度卷积特征的图像标注方法研究[D].杭州:浙江大学,2017.Song G.Image annotation method based on transfer learning and deep convolutional feature[D].Hangzhou:Zhejiang University,2017.