基于卷积神经网络的水稻纹枯病图像识别

doi:10.16819/j.1001-7216.2019.8051

摘要/Abstract

摘要：

【目的】水稻纹枯病是影响水稻生产的三大病害之一。研究卷积神经网络对水稻纹枯病的自动识别,弥补人工识别的不足,对预防和准确识别水稻蚊枯病类型有着重要意义。【方法】以卷积神经网络进行水稻纹枯病识别,并与基于支持向量机的识别方法进行对比。【结果】卷积神经网络识别率达到97%,优于支持向量机的95%。【结论】卷积神经网络运用于水稻纹枯病识别是可行的,弥补了人工识别的不足。此算法训练的模型有着较好的识别性能。

关键词: 水稻纹枯病, 卷积神经网络, 分类识别

Abstract:

【Objective】Rice sheath blight is one of the three major diseases in rice production.The convolutional neural network which stands out for automatic identification of rice shealth blight can compensate for the lack of human identification. To solve this problem and prevent diseases deterioration, accurate identification of diseases types is of great significance.【Method】The convolutional neural network method was used to recognize rice sheath blight and compared with the recognition method based on support vector machine.【Result】The convolutional neural network method showed the recognition rate of 97%, better than that of support vector machine(95%).【Conclusion】The application of convolutional neural network to the identification of rice sheath blight is feasible and makes up for the lack of artificial recognition. The model trained by this algorithm has great recognition performance.

Key words: Rhizocotonia solani, convolutional neural network, classification and recognition

中图分类号:

刘婷婷, 王婷, 胡林. 基于卷积神经网络的水稻纹枯病图像识别[J]. 中国水稻科学, 2019, 33(1): 90-94.

Tingting LIU, Ting WANG, Lin HU. Rhizocotonia Solani Recognition Algorithm Based on Convolutional Neural Network[J]. Chinese Journal OF Rice Science, 2019, 33(1): 90-94.

图/表 3

参考文献 41

[1]	桑海旭, 王井士, 刘郁, 李运动, 马晓慧, 刘志恒. 水稻纹枯病对水稻产量及米质的影响. 北方水稻, 2013, 43(1): 10-13.
	Sang H X, Wang J S, Liu Y,Li Y D, Ma X H, Liu Z H.The effect of rice sheath blight on yield and quality.North Rice, 2013, 43(1): 10-13. (in Chinese)
[2]	李雪婷, 徐梦亚, 郑少兵, 孙正祥, 周燚. 水稻纹枯病研究进展. 长江大学学报: 自然科学版, 2017, 14(14): 15-18.
	Li X T, Xu M Y, Zheng S B, Sun Z X, Zhou Y.Research progress in rice sheath blight.J Yangtze Univ: Nat Sci Ed, 2017, 14(14): 15-18. (in Chinese with English abstract).
[3]	路桂珍, 杨秀军. 应用红外遥感技术监测植物病虫害.红外技术, 1990, 12(2): 18-21.
	Lu G Z, Yang X J.Application of infrared remote sensing technology in monitoring plant diseases and insect pests.Infrar Technol, 1990, 12(2): 18-21. (in Chinese with English abstract).
[4]	Sasaki Y, Kamoto T O.Automatic diagnosis of plant disease.J JSAM, 1999, 61(2): 119-126.
[5]	Gassoumi H, Prasad N R, Ellington J J.Neural network-based approach for insect classification in cotton ecosystems. Bangkok, Thailand: International Conference on Intelligent Technologies (InTech 2000), 2000, 7.
[6]	陈佳娟, 赵学笃. 采用计算机视觉进行棉花虫害程度的自动测试. 农业工程报, 2001, 17(2): 157-160.
	Chen J J, Zhao X D.Automatic measurement of danger degree of cotton insect pests using computer vision.Transa Chin Soc Agric Engin, 2001, 17(2): 157-160. (in Chinese with English abstract)
[7]	徐贵力, 毛罕平, 李萍萍. 差分百分率直方图法提取缺素叶片纹理特征. 农业机械学报, 2003, 34(2): 76-79.
	Xu G L, Mao H P, Li P P.Research on extraction leaf texture features as sample of nutrient shortage by percent histogram of differentiation.Transac Chin Soc Agric Machin, 2003, 34(2): 76-79.(in Chinese with English abstract)
[8]	田有文, 张长水, 李成华. 基于支持向量机和色度矩的植物病害识别研究. 农业机械学报, 2004, 35(3): 95-98.
	Tian Y W, Zhang C S, Li C H.Study on plant disease recognition using support vector machine and chromaticity moments.Transac Chin Soc Agric Mach, 2004, 35(3): 95-98. (in Chinese with English abstract)
[9]	田有文, 李成华. 基于图像处理的日光温室黄瓜病害识别的研究. 农机化研究, 2006(2): 151-153.
	Tian Y W, Li C H.Research on recognition of cucumber disease based on image processing in sunlight greenhouse J Agric Mech Res, 2006(2): 151-153. (in Chinese with English abstract)
[10]	张建华, 祁力钧, 冀荣华, 王虎, 黄士凯, 王沛. 基于粗糙集和BP神经网络的棉花病害识别. 农业工程学报, 2012, 28(7): 161-167.
	Zhang J H, Qi L J, Ji R H, Wang H, Huang S K, Wang P.Cotton diseases identification based on rough sets and BP neural network.Tran Chin Soc Agric Engin, 2012, 28(7): 161-167. (in Chinese with English abstract)
[11]	王献锋, 张善文, 王震, 张强. 基于叶片图像和环境信息的黄瓜病害识别方法.农业工程学报, 2014, 30(14): 148-153.
	Wang X F, Zhang S W, Wang Z, Zhang Q.Recognition of cucumber diseases based on leaf image and environmental information.Transac Chin Soc Agric Engin, 2014, 30(14): 148-153. (in Chinese with English abstract)
[12]	Rumelhart D E, Hinton G E, Williams R J.Learning internal representations by error propagation. DTIC Document, 1985.
[13]	Lecun Y, Boser B, Denker J S, Henderson D, Howard R E, Hubbard W, Jackel L D.Backprogation applied to handwritten zip code recognition.Neural Comput, 1989, 1(4): 541-551.
[14]	Lecun Y, Bottou L, Bengio Y, Haffner P.Gradient-based learning applied to document recognition.Proc IEEE, 1998, 86(11): 2278-2324.
[15]	Abhishek S, Nasser K.A convolutional neural network smartphone app for real-time voice activity detection.IEEE Access, 2018, 6: 9017-9026.
[16]	Wang P, Xu B, Xu J M, Tian G H, Liu C L, Hao H W.Semantic expansion using word embedding clustering and convolutional neural network for improving short text classification.Neurocomputing, 2016, 174: 806-814.
[17]	Zhao M, Li H, Shi X, Chan Y, Luo X, Li T.Automated recognition of zygote cytoplasmic area (ZCA) in time-lapse imaging (TLI) based on deep convolutional neural network (CNN).Fertil Steril, 2017, 108(3): 239.
[18]	Tung T S, Lee G.Language identification in handwritten words using a convolutional neural network.Int J Contents, 2017, 13(3): 38-42.
[19]	Shi B G, Bai X, Yao C.Script identification in the wild via discriminative convolutional neural network.Pat Recog, 2016, 52: 448-458.
[20]	Syafeeza A R, Mohamed K H, Rabia B.Finger-vein biometric identification using convolutional neural network.Turkish J Elect Engineer Compu Sci, 2016, 24: 1863-1878.
[21]	Peng M, Wang C Y, Chen T, Liu G Y.NIRFaceNet: A convolutional neural network for near-infrared face identification.Information, 2016, 7(4): 61.
[22]	Jorden H, Victoria L, Vit G K, Purang A, Robert R.SLIDE: automatic spine level identification system using a deep convolutional neural network.Intern J Com Assisted Radiol Surg, 2017, 12(7): 1189-1198.
[23]	Zhang W, Du Y H, Taketoshi Y, Wang Q.DRI-RCNN: An approach to deceptive review identification using recurrent convolutional neural network.Informat Process Manag, 2018, 54(4): 576-592.
[24]	Lee S J, Yun J P, Koo G, Kim S W.End-to-end recognition of slab identification numbers using a deep convolutional neural network.Knowledge-Based Sys, 2017, 132: 1-10.
[25]	Zhang Y Y, Zhao D, Sun J D, Zou G F, Li W T.Adaptive convolutional neural network and its application in face recognition.Neural Process Let, 2016, 43(2): 389-399.
[26]	Lecun Y, Bottou L, Bengio Y, Haffner P.Gradient-based learning applied to document recognition.Proc IEEE, 1998, 86(11): 2278-2324.
[27]	Krizhevsky A, Sutskever I, Hinton G E.Imagenet classification with deep convolutional neural networks. Proceedings of the Advances in Nerural Information Processing System. South Lake Tahoe,US: 2012: 1097-1105.
[28]	Han X B, Zhong Y F, Cao L Q, Zhang L P.Pre-trained alexnet architecture with pyramid pooling and supervision for high spatial resolution remote sensing image scene classification.Remote Sens, 2017, 9(8): 848.
[29]	Szegedy C, Liu W, Jia Y Q, Sermanet P, Reed S, Anguelov D, Erhan D, Vanhoucke V, Rabinovich A. Going deeper with convolutions. Boston MA,USA:2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2015.
[30]	Mostafa M G, Berrin Y, Erchan A.Plant identification using deep neural networks via optimization of transfer learning parameters.Neurocomputing, 2017, 235: 228-235.
[31]	Simonyan Karen,Zisserman. Very Deep Convolutional Networks for Large-Scale Image Recognition.ICLR, 2015.
[32]	Wang H Y, He Z S, Huang Y W, Chen D D, Zhou Z X.Bodhisattva head images modeling style recognition of Dazu Rock Carvings based on deep convolutional network.J Cul Herit, 2017, 27: 60-71.
[33]	He K M, Zhang X Y, Ren S Q, Sun J.Deep residual learning for image recognition. Las Vegas, NV, USA:2016 IEEE Conference on Computer Vision and Pattern Recognition(CVPR), 2016: 770-778.
[34]	Sun W F, Yao B, Chen B Q, He Y C, Cao X C, Zhou T X, Liu H G.Noncontact surface roughness estimation using 2D complex wavelet enhanced resnet for intelligent evaluation of milled metal surface quality.Appl Sci, 2018, 8(3): 381.
[35]	Pang Y W, Sun M L, Jiang X H, Li X L.Convolution in Convolution for Network in Network.IEEE Trans Neural Networks Learning Syst, 2018, 29(5): 1587-1597.
[36]	Chen Y, Yang X N, Zhong B N, Zhang H Z, Lin C L.Network in network based weakly supervised learning for visual tracking.J Vis Comm Image Repres, 2016, 37: 3-13.
[37]	Vladimir N,.Vapnik. Statistical learning theory. A Wiley- Interscience Publication John wiley &son INC, 1998.
[38]	顾亚祥, 丁世飞. 支持向量机研究进展. 计算机科学, 2011, 38(2): 14-17.
	Gu Y X, Ding S F.Advances of support vector machines(SVM).Com Sci, 2011, 38(2): 14-17. (in Chinese with English abstract)
[39]	Gonzalez R C, Woods R E. Digital Image Processing.4th edition. Pearson, 2017.
[40]	Gotlieb C C, Kreyszig H E.Texture descriptors based on co-occurrence matrices.Com Vision Grap & Image Proc, 1990, 51(1): 70-86.
[41]	毋媛媛, 刁智华, 王会丹, 赵明珍, 魏伟. 作物病害图像形状特征提取研究. 农机化研究, 2015, 37(1): 64-67.
	Wu Y Y, Diao Z H, Wang H D, Zhao M Z, Wei W.Research on shape feature extraction of crop disease image.J Agric Mec Res, 2015, 37(1): 64-67. (in Chinese)

样本正确识别率 Correct recognition rate of samples	第一次 First	第二次 Second	第三次 Third	第四次 Fourth	第五次 Fifth
卷积神经网络CNN	100.0	95.0	92.5	97.5	100.0
支持向量机SVM	97.5	92.5	90.0	95.0	100.0

样本正确识别率 Correct recognition rate of samples	第一次 First	第二次 Second	第三次 Third	第四次 Fourth	第五次 Fifth
卷积神经网络CNN	100.0	95.0	92.5	97.5	100.0
支持向量机SVM	97.5	92.5	90.0	95.0	100.0

识别算法 Recognition algorithm	观测数 Observation number	求和 Sum	平均 Average	方差 Variance
行1：卷积神经网络Row1:CNN	5	4.850	0.9700	0.0011
行2：支持向量机Row2:SVM	5	4.750	0.9500	0.0016
列1：第一次Column1:First	2	1.975	0.9875	0.0003
列2：第二次Column2:Second	2	1.875	0.9375	0.0003
列3：第三次Column3:Third	2	1.825	0.9125	0.0003
列4：第四次Column4:Fourth	2	1.925	0.9625	0.0003
列5：第五次Column5:Fifth	2	2.000	1.0000	0.0003

识别算法 Recognition algorithm	观测数 Observation number	求和 Sum	平均 Average	方差 Variance
行1：卷积神经网络Row1:CNN	5	4.850	0.9700	0.0011
行2：支持向量机Row2:SVM	5	4.750	0.9500	0.0016
列1：第一次Column1:First	2	1.975	0.9875	0.0003
列2：第二次Column2:Second	2	1.875	0.9375	0.0003
列3：第三次Column3:Third	2	1.825	0.9125	0.0003
列4：第四次Column4:Fourth	2	1.925	0.9625	0.0003
列5：第五次Column5:Fifth	2	2.000	1.0000	0.0003