堆栈降噪自编码结合随机森林的黄龙病检测
|
摘要
近红外光谱分析技术作为一种无损、快捷的分析方法在各个领域应用相当广泛。针对柑橘黄龙病检测成本高、可靠性差和精度低等问题,提出了一种堆栈降噪自编码融合随机森林(Stacked Denoising Auto-encoders Combined Random Forest,SDAE-RF)的柑橘黄龙病近红外光谱检测方法,该方法首先采用多阶段预处理法对样本光谱数据进行预处理,然后采用SDAE对经过预处理后的光谱数据进行降维,实现柑橘样本深层特征的提取,最后利用RF的投票集成策略实现分类鉴别。为了验证SDAE-RF模型的性能,采用某公司提供的柑橘叶片近红外光谱数据为实例,以不同比例的训练集进行实验,并与ELM、SWELM、SVM、BP、SDAE和RF模型的鉴别能力进行对比。实验结果表明,SDAE-RF模型较其他算法在分类精度、算法稳定性以及训练时间方面均表现出较好的效果。
Near-infrared spectroscopy was widely used as a non-destructive and fast method in agriculture.To solve the low accuracy,poor reliability and high cost of the detection of Huanglongbing,amethod of citrus near-infrared spectroscopy for Stacking Denoising Auto-encoders Combined Random Forestis proposed.The method firstly preprocesses the spectral data by multi-stage preprocessing.Then,by usingSDAEthedimensionof the pre-processed spectral datasetare reduced and deep features of the sample are extracted.Finally,the voting ensemble strategy of RF is used to realize classification and identification.In order to verify the performance of the SDAE-RF model,the near-infrared spectrum data of citrus leaves provided by a company were taken as an example.Experiments were carried out with different proportions of training sets,and compared with the identification capabilities of ELM,SWELM,SVM,BP,SDAE and RF models.Experimental results show that the SDAE-RF model performs better than other algorithms in terms of classification accuracy,algorithm stability and train time.
Near-infrared spectroscopy was widely used as a non-destructive and fast method in agriculture.To solve the low accuracy,poor reliability and high cost of the detection of Huanglongbing,amethod of citrus near-infrared spectroscopy for Stacking Denoising Auto-encoders Combined Random Forestis proposed.The method firstly preprocesses the spectral data by multi-stage preprocessing.Then,by usingSDAEthedimensionof the pre-processed spectral datasetare reduced and deep features of the sample are extracted.Finally,the voting ensemble strategy of RF is used to realize classification and identification.In order to verify the performance of the SDAE-RF model,the near-infrared spectrum data of citrus leaves provided by a company were taken as an example.Experiments were carried out with different proportions of training sets,and compared with the identification capabilities of ELM,SWELM,SVM,BP,SDAE and RF models.Experimental results show that the SDAE-RF model performs better than other algorithms in terms of classification accuracy,algorithm stability and train time.
引文
[1] Ríos-Reina R,García-González D L,Callejón R M,et al.NIR spectroscopy and chemometrics for the typification of Spanish wine vinegars with a protected designation of origin[J].Food Control,2018,89.
[2] ZHAO Ke,XIONG Yan,ZHAO Ming.Rapid non-destructive testing of navel orange based on near infrared spectroscopy[J].Laser & Infrared,2011,41(6):649-652.(in Chinese)赵珂,熊艳,赵敏.基于近红外光谱技术的脐橙快速无损检测[J].激光与红外,2011,41(6):649-652.
[3] Sankaran S,Ehsani R.Visible-near infrared spectroscopy based citrus greening detection:evaluation of spectral feature extraction techniques[J].Crop Protection,2011,30(11):1508-1513.
[4] Wetterich C B,Ruan F D ON,Belasque J,et al.Detection of citrus canker and Huanglongbing using fluorescence imaging spectroscopy and support vector machine technique[J].Appl Opt,2016,55(2):400-407.
[5] Sankaran S,Maja J M,Buchanon S,et al.Huanglongbing(citrus greening) detection using visible,near infrared and thermal imaging techniques[J].Sensors,2013,13(2):2117-2130.
[6] Li X H,Li M Z,Won S L,et al.Visible-NIR spectral feature of citrus greening disease[J].Spectroscopy and Spectral Analysis,2014,34(6):1553-1559.
[7] DENG Xiaoling,KONG Chen,DENG XiaoLing.Detection of citrus huanglongbing based on principal component analysis and back propagation neural network[J].Acta Photonica Sinica,2014,43(4):430002-0430002.(in Chinese)邓小玲,孔晨,邓晓玲.基于主成分分析和BP神经网络的柑橘黄龙病诊断技术[J].光子学报,2014,43(4):430002-0430002.
[8] SUN Xudong,LIU Yande,XIAO Huaichun,et al.Rapid diagnosis of sound,yellow and citrus greening leaves with hyperspectral imaging[J].Spectroscopy and Spectral Analysis,2017,37(2):551-556.(in Chinese)孙旭东,刘燕德,肖怀春,等.正常、缺素和黄龙病柑桔叶片高光谱成像快速诊断[J].光谱学与光谱分析,2017,37(2):551-556.
[9] LIU Yande,XIAO Huaichun,SUN Xudong,et al.Study on the quick non-destructive detection of citrus huanglongbing based on the spectrometry of VIS and NIR[J].Spectroscopy and Spectral Analysis,2018(2):528-534.(in Chinese)刘燕德,肖怀春,孙旭东,等.基于可见与近红外光谱联用的柑桔黄龙病快速无损检测研究[J].光谱学与光谱分析,2018(2):528-534.
[10] Donahue J,Hendricks L A,Rohrbach M,et al.Long-term recurrent convolutional networks for visual recognition and description[J].IEEE Trans.patt.Anal.Mach.Intell.,2017,39(4):677-691.
[11] LIU Zhixiao,LIANG Liang,YU Xiaoying.Identification of the types of waste water based on visible/near-infrared spectroscopy and BP-ANN algorithm[J].Laser & Infrared,2009,39(11):1153-1157.(in Chinese)刘志霄,梁亮,俞晓莹.基于可见-近红外光谱技术与BP-ANN算法的污水类型鉴定[J].激光与红外,2009,39(11):1153-1157.
[12] Akhtar N,Mian A.Threat of adversarial attacks on deep learning in computer vision:a survey[J].IEEE Access,2018,6:14410-14430.
[13] Zhang Z,Luo P,Chen C L,et al.Learning deep representation for face alignment with auxiliary attributes[J].IEEE Transactions on Pattern Analysis & Machine Intelligence,2016,38(5):918-930.
[14] Vincent P,Larochelle H,Lajoie I,et al.Stacked denoising autoencoders:learning useful representations in a deep network with a local denoising criterion[J].Journal of Machine Learning Research,2010,11(12):3371-3408.
[15] Surhone L M,Tennoe M T,Henssonow S F,et al.Random Forest[J].Machine Learning,2010,45(1):5-32.
[2] ZHAO Ke,XIONG Yan,ZHAO Ming.Rapid non-destructive testing of navel orange based on near infrared spectroscopy[J].Laser & Infrared,2011,41(6):649-652.(in Chinese)赵珂,熊艳,赵敏.基于近红外光谱技术的脐橙快速无损检测[J].激光与红外,2011,41(6):649-652.
[3] Sankaran S,Ehsani R.Visible-near infrared spectroscopy based citrus greening detection:evaluation of spectral feature extraction techniques[J].Crop Protection,2011,30(11):1508-1513.
[4] Wetterich C B,Ruan F D ON,Belasque J,et al.Detection of citrus canker and Huanglongbing using fluorescence imaging spectroscopy and support vector machine technique[J].Appl Opt,2016,55(2):400-407.
[5] Sankaran S,Maja J M,Buchanon S,et al.Huanglongbing(citrus greening) detection using visible,near infrared and thermal imaging techniques[J].Sensors,2013,13(2):2117-2130.
[6] Li X H,Li M Z,Won S L,et al.Visible-NIR spectral feature of citrus greening disease[J].Spectroscopy and Spectral Analysis,2014,34(6):1553-1559.
[7] DENG Xiaoling,KONG Chen,DENG XiaoLing.Detection of citrus huanglongbing based on principal component analysis and back propagation neural network[J].Acta Photonica Sinica,2014,43(4):430002-0430002.(in Chinese)邓小玲,孔晨,邓晓玲.基于主成分分析和BP神经网络的柑橘黄龙病诊断技术[J].光子学报,2014,43(4):430002-0430002.
[8] SUN Xudong,LIU Yande,XIAO Huaichun,et al.Rapid diagnosis of sound,yellow and citrus greening leaves with hyperspectral imaging[J].Spectroscopy and Spectral Analysis,2017,37(2):551-556.(in Chinese)孙旭东,刘燕德,肖怀春,等.正常、缺素和黄龙病柑桔叶片高光谱成像快速诊断[J].光谱学与光谱分析,2017,37(2):551-556.
[9] LIU Yande,XIAO Huaichun,SUN Xudong,et al.Study on the quick non-destructive detection of citrus huanglongbing based on the spectrometry of VIS and NIR[J].Spectroscopy and Spectral Analysis,2018(2):528-534.(in Chinese)刘燕德,肖怀春,孙旭东,等.基于可见与近红外光谱联用的柑桔黄龙病快速无损检测研究[J].光谱学与光谱分析,2018(2):528-534.
[10] Donahue J,Hendricks L A,Rohrbach M,et al.Long-term recurrent convolutional networks for visual recognition and description[J].IEEE Trans.patt.Anal.Mach.Intell.,2017,39(4):677-691.
[11] LIU Zhixiao,LIANG Liang,YU Xiaoying.Identification of the types of waste water based on visible/near-infrared spectroscopy and BP-ANN algorithm[J].Laser & Infrared,2009,39(11):1153-1157.(in Chinese)刘志霄,梁亮,俞晓莹.基于可见-近红外光谱技术与BP-ANN算法的污水类型鉴定[J].激光与红外,2009,39(11):1153-1157.
[12] Akhtar N,Mian A.Threat of adversarial attacks on deep learning in computer vision:a survey[J].IEEE Access,2018,6:14410-14430.
[13] Zhang Z,Luo P,Chen C L,et al.Learning deep representation for face alignment with auxiliary attributes[J].IEEE Transactions on Pattern Analysis & Machine Intelligence,2016,38(5):918-930.
[14] Vincent P,Larochelle H,Lajoie I,et al.Stacked denoising autoencoders:learning useful representations in a deep network with a local denoising criterion[J].Journal of Machine Learning Research,2010,11(12):3371-3408.
[15] Surhone L M,Tennoe M T,Henssonow S F,et al.Random Forest[J].Machine Learning,2010,45(1):5-32.