基于可见近红外光谱比较主成分回归、偏最小二乘回归和反向传播神经网络对土壤氮的预测研究

李硕; 汪善勤; 张美琴

doi:doi:10.3788/aos201232.0830001

光学学报, 2012, 32 (8): 0830001, 网络出版: 2012-07-02

基于可见近红外光谱比较主成分回归、偏最小二乘回归和反向传播神经网络对土壤氮的预测研究

Comparison Among Principal Component Regression, Partial Least Squares Regression and Back Propagation Neural Network for Prediction of Soil Nitrogen with Visible-Near Infrared Spectroscopy

论文大纲

李硕 ^1,*汪善勤 ^1,2张美琴 ¹

作者单位

¹ 华中农业大学资源与环境学院, 湖北武汉 430070

² 中国科学院土壤与农业可持续发展国家重点实验室, 江苏南京 210008

摘要

建模方法是影响可见近红外光谱定量结果的主要因素之一。在470~1000 nm波段的12个土壤剖面对48个剖面样经过风干、研磨、过筛后进行光谱采集。经一阶微分变换及Savizky-Golay平滑处理后，分别应用主成分回归(PCR)、偏最小二乘回归(PLSR)和反向传播神经网络(BPNN)3种方法建立土壤全氮(TN)的定量模型。 PCR与PLSR两线性模型的决定系数(R2)分别为0.74和0.8，其剩余预测偏差(RPD)分别为2.23和2.22，但两模型仅能用于TN的粗略估计。由PCR提供主成分数，PLSR提供潜变量(LV)数分别作为BPNN的输入所构建的两个非线性模型均明显优于线性模型PCR和PLSR。其中以4个LV作为输入的BPNN-LV模型预测性能最优，R2 以及RPD分别达到0.9和3.11。实验结果表明，提取可见近红外光谱的PLSR LV因子作为BPNN的输入，所建定量模型可用于土壤氮纵向时空分布的快速准确预测。

Abstract

The selection of modeling method is one of the main factors influencing the quantitative accuracy with visible-near infrared (Vis-NIR) spectroscopy. We compare the performance of three calibrations methods, i.e., principal component regression (PCR), partial least squares regression (PLSR), and back propagation neural network (BPNN) based on Vis-NIR reflectance spectra of soil total nitrogen (TN) quantitative forecast results. Covered in the 470~1000 nm wavelength range, spectroscopy of 48 soil samples selected from 12 profiles are air-dried, screened and mushed, then processed by the first order derivative and Savizky-Golay smoothing methods. Leave-one-out cross validation is also adopted to determine the optimal factor numbers. The results indicate that PCR and PLSR linear models are able to meet general prediction and with little difference, where coefficients of determination (R2) are 0.74 and 0.8, respectively, and residual predictive deviation (RPDs) are 2.23 and 2.22. The two nonlinear models built by BPNN in combination with PCR and PLSR, respectively, are superior to the linear models of PCR and PLSR in the precision of prediction. BPNN, principal components (PCs) whose input is the PCs resulted from the PCR, while the BPNN latent variables (LVs) whose input is the first 4 LV results obtained from PLSR has the best performance (R2=0.9, RPD is3.11). It is recommended to adopt BPNN-LV model to rapidly predict the vertical spatial and temporal distribution of TN with Vis-NIR spectroscopy.

参考文献

[1] 孙光明, 刘飞, 张帆等. 基于近红外光谱技术检测除草剂胁迫下油菜叶片中脯氨酸含量的方法［J］. 光学学报, 2010, 30(4): 1192～1196

Sun Guangming, Liu Fei, Zhang Fan et al.. Fast determination of proline in herbicide-stressed oilseed rape leaves based on near infrared spectroscopy［J］. Acta Optica Sinica, 2010, 30(4): 1192～1196

[2] 郭伟良, 王丹, 宋佳等. 近红外光谱法同时快速定量分析蛹虫草菌丝体中4种有效成分［J］. 光学学报, 2011, 31(2): 0230002

Guo Weiliang, Wang Dan, Song Jia et al.. Simultaneous and rapid quantitative analysis of four components in Cordyceps militaris mycelium powder using near infrared spectroscopy［J］. Acta Optica Sinica, 2011, 31(2): 0230002

[3] 刘燕德, 陈兴苗, 欧阳爱国. 可见/近红外光谱法无损检测赣南脐橙可溶性固形物［J］. 光学学报, 2008, 28(3): 478～481

Liu Yande, Chen Xingmiao, Ouyang Aiguo. Non-destructive measurement of soluble solid content in Gannan navel oranges by visible/near-infrared spectroscopy［J］. Acta Optica Sinica, 2008, 28(3): 478～481

[4] C.-W. Chang, D. A. Laird, M. J. Mausbach et al.. Near-infrared reflectance spectroscopy: principal components regression analyses of soil properties［J］. Soil Science Society of America Journal, 2001, 65(2): 480～490

[5] B. H. Kusumo, M. J. Hedley, C. B. Hedley et al.. Measuring carbon dynamics in field soils using soil spectral reflectance: prediction of maize root density, soil organic carbon and nitrogen content［J］. Plant and Soil, 2011, 338(1-2): 233～245

[6] 郑立华, 李民赞, 潘娈等. 基于近红外光谱技术的土壤参数BP神经网络预测［J］. 光谱学与光谱分析, 2008, 28(5): 1160～1164

Zheng Lihua, Li Minzan, Pan Luan et al.. Estimation of soil organic matter and soil total nitrogen based on NIR spectroscopy and BP neural network［J］. Spectroscopy and Spectral Analysis, 2008, 28(5): 1160～1164

[7] 高英志, 汪诗平, 韩兴国等. 退化草地恢复过程中土壤氮素状况以及与植被地上绿色生物量形成关系的研究［J］. 植物生态学报, 2004, 28(3): 285～293

Gao Yingzhi, Wang Shiping, Han Xingguo et al.. Soil nitrogen regime and the relationship between aboveground green phytobiomass and soil nitrogen fractions at different stocking rates in the Xilin River Basin, Inner Mongolia [J]. Acta Phytoecologica Sinica, 2004, 28(3): 285～293

[8] 路鹏, 魏志强, 牛铮. 应用特征波段和反射变形差的方法进行土壤属性估算［J］. 光谱学与光谱分析, 2009, 29(3): 716～721

Lu Peng, Wei Zhiqiang, Niu Zheng. Estimate of soil attributes using the method of special band and reflectance inflection difference［J］. Spectroscopy and Spectral Analysis, 2009, 29(3): 716～721

[9] 卢艳丽, 白由路, 王磊等. 黑土土壤中全氮含量的高光谱预测分析［J］. 农业工程学报, 2010, 26(1): 256～261

Lu Yanli, Bai Youlu, Wang Lei et al.. Determination for total nitrogen content in black soil using hyperspectral data［J］. Transactions of the Chinese Society of Agricultural Engineering, 2010, 26(1): 256～261

[10] 徐永明, 蔺启忠, 黄秀华等. 利用可见光/近红外反射光谱估算土壤总氮含量的实验研究［J］. 地理与地理信息科学, 2005, 21(1): 19～22

Xu Yongming, Lin Qizhong, Huang Xiuhua et al.. Experimental study on total nitrogen concentration in soil by VNIR reflectance spectrum［J］. Geography and Geo-Information Science, 2005, 21(1): 19～22

[11] R.A. V. Rossel, D. J. J. Walvoort, A. B. McBratney et al.. Visible, near infrared, mid infrared or combined diffuse reflectance spectroscopy for simultaneous assessment of various soil properties［J］. Geoderma, 2006, 131(1-2): 59～75

[12] 高云, 汪善勤, 李小昱等. 三轴向土壤高光谱成像实验台: 中国, 200920087238.7 [P]. 2010-07-14

[13] 褚小立, 许育鹏, 陆婉珍. 偏最小二乘法方法在光谱定性分析中的应用研究［J］. 现代仪器, 2007, (5): 13～15

[14] L. J. Janik, D. Cozzolino, R. Dambergs et al.. The prediction of total anthocyanin concentration in red-grape homogenates using visible-near-infrared spectroscopy and artificial neural networks［J］. Analytica Chimica Acta, 2007, 594(1): 107～118

[15] R. Zornoza, C. Guerrero, J. Mataix-Solera et al.. Near infrared spectroscopy for determination of various physical, chemical and biochemical properties in Mediterranean soils［J］. Soil Biology & Biochemistry, 2008, 40(7): 1923～1930

[16] D. Cozzolino, A. Moron. Potential of near-infrared reflectance spectroscopy and chemometrics to predict soil organic carbon fractions［J］. Soil & Tillage Research, 2006, 85(1-2): 78～85

[17] C. S. T. Daughtry. Discriminating crop residues from soil by shortwave infrared reflectance［J］. Agronomy Journal, 2001, 93(1): 125～131

[18] K. P. Fabrizzi, A. Moron, F. O. Garcia. Soil carbon and nitrogen organic fractions in degraded versus non-degraded mollisols in Argentina［J］. Soil Science Society of America Journal, 2003, 67(6): 1831～1841

[19] M. Zimmermann, J. Leifeld, J. Fuhrer. Quantifying soil organic carbon fractions by infrared-spectroscopy［J］. Soil Biology & Biochemistry, 2007, 39(1): 224～231

[20] A. M. Mouazen, B. Kuang, J. De Baerdemaeker et al.. Comparison among principal component, partial least squares and back propagation neural network analyses for accuracy of measurement of selected soil properties with visible and near infrared spectroscopy［J］. Geoderma, 2010, 158(1-2): 23～31

李硕, 汪善勤, 张美琴. 基于可见近红外光谱比较主成分回归、偏最小二乘回归和反向传播神经网络对土壤氮的预测研究[J]. 光学学报, 2012, 32(8): 0830001. Li Shuo, Wang Shanqin, Zhang Meiqin. Comparison Among Principal Component Regression, Partial Least Squares Regression and Back Propagation Neural Network for Prediction of Soil Nitrogen with Visible-Near Infrared Spectroscopy[J]. Acta Optica Sinica, 2012, 32(8): 0830001.

基于可见近红外光谱比较主成分回归、偏最小二乘回归和反向传播神经网络对土壤氮的预测研究

关于本站 Cookie 的使用提示

全站搜索

基于可见近红外光谱比较主成分回归、偏最小二乘回归和反向传播神经网络对土壤氮的预测研究

相关论文

相关资讯

关于本站 Cookie 的使用提示

全站搜索