变换光谱数据对土壤氮素PLSR模型的影响研究

doi:10.3724/SP.J.1047.2016.01123

摘要/Abstract

摘要：

光谱数据变换对消除背景、噪音影响以及提取光谱特征有重要的作用,是光谱数据分析过程中的必要步骤。为了研究光谱变换处理对土壤氮素PLSR模型的影响精度,并选择最佳光谱变换处理方法,本文对原始光谱数据进行了15种典型光谱变换,通过比较不同变换光谱与土壤氮素的相关性,实现土壤氮素的PLSR精确诊断,并综合评定最佳光谱数据变换方法。结果表明,涉及微分处理后的光谱变换,尤其是先进行开方（T₈、T₁₁）、对数（T₆、T₁₂）等变换后再进行微分处理,可提高其与土壤氮素的相关性。在引入较少因子变量个数的条件下,该方法使因变量解释量达到了98%。综合考虑模型的校正、验证效果及模型复杂度（模型最佳因子变量个数）,可得出光谱平方根的一阶微分变换处理（T₈）为最佳的土壤光谱变换算法。该条件下的土壤氮素的校正模型表现为R²=0.985、RMSE_C=0.000132、F_n=6,验证模型的表现为R²=0.9853、RMSE_V=0.000162,结果表明基于T₈的光谱数据变换可实现本试验条件下土壤氮素的光谱估算。另外,可以考虑将原始光谱的一阶微分（T₉）、对数和对数倒数的一阶微分（T₆、T₇）以及平方根和对数的二阶微分（T₁₁、T₁₂）作为光谱数据变换方法。本文研究结果可为土壤氮素估算和光谱数据预处理提供技术参考。

关键词: 光谱变换, 土壤氮素, 诊断, 精度, 偏最小二乘法

Abstract:

Spectral transformation is an essential pre-treatment technique as it can eliminate the effects of background and noise, and it plays a vital role in extracting spectral features′ information and constructing the optimal model. In order to explore the effects of different spectral transformation methods on the accuracy of the PLSR model in monitoring soil nitrogen and determining the optimal spectral transformation, the raw spectrum was transformed with respect to fifteen transformation algorithms and the correlations between each pair of transformed spectrum and soil nitrogen were analyzed. Furthermore, the performances of the PLSR models in monitoring the soil nitrogen based on different transformed spectra were evaluated. The results showed that, for cases involving the first or second-order differential reprocessing transformations, the correlation coefficient between the soil nitrogen and the relevant transformed spectrum increased more significantly than with the raw spectrum, especially when applying the transformation algorithms of square root (T₈ and T₁₁) and logarithm (T₆ and T₁₂) firstly. Also, fewer optimal factors for these pre-treatments were needed and selected to achieve the threshold of 98% in explaining the dependent variable. Moreover, the first-order differential reprocessing of the square root of raw spectrum (T₈) had a higher accuracy (R²=0.985022, RMSEC=0.000132; R²=0.9853, RMSEV=0.000162, F_n=6) for the calibrated model and the validated model respectively, after the comprehensive evaluation of the predicting performance and the complexity of different models. Finally, the first-order differential reprocessing of the square root of raw spectrum (T₈) was determined as the recommended transformation method to evaluate the soil nitrogen. In addition, the first-order and second-order differential of the logarithm of raw spectrum (T₆ and T₁₂), the first-order differential of the logarithmic reciprocal of raw spectrum (T₇), the first-order differential of raw spectrum (T₉), as well as the second-order differential of the square root of raw spectrum (T₁₁) could also be considered and chosen as alternatives. The study would provide some theoretical techniques and references to the evaluation of soil nitrogen and spectrum processing.

Key words: spectral transformation, soil nitrogen, diagnosis, accuracy, PLSR

乔星星, 冯美臣, 杨武德, 孙慧, 郭小丽, 史超超. 变换光谱数据对土壤氮素PLSR模型的影响研究[J]. 地球信息科学学报, 2016, 18(8): 1123-1132.DOI:10.3724/SP.J.1047.2016.01123

QIAO Xingxing,FENG Meichen,YANG Wude,SUN Hui,GUO Xiaoli,SHI Chaochao. Effect of Spectral Transformation Processes on the PLSR Models of Soil Nitrogen[J]. Journal of Geo-information Science, 2016, 18(8): 1123-1132.DOI:10.3724/SP.J.1047.2016.01123

图/表 7

表1

16种光谱数据变换算法信息"

简写	变换算法	简写	变换算法
T₀	R	T₈	$R'$
T₁	$1 / R$	T₉	$R'$
T₂	$lgR$	T₁₀	$1 / lgR ″$
T₃	$1 / lgR$	T₁₁	$R ″$
T₄	$R$	T₁₂	$(lgR) ″$
T₅	$1 R'$	T₁₃	$R 2$
T₆	$(lgR)'$	T₁₄	$R 3$
T₇	$(1 / lgR)'$	T₁₅	$(lg R 2)'$

表1

表2

图1

图2

图3

图4

表3

参考文献 31

[1]	Zhao S C, Qiu S J, Cao C Y, et al.Responses of soil properties, microbial community and crop yields to various rates of nitrogen fertilization in a wheat-maize cropping system in north-central China[J]. Agriculture, Ecosystems and Environment, 2014,194:29-37. doi: 10.1016/j.agee.2014.05.006
[2]	Gnyp M L, Miao Y X, Yuan F, et al.Hyperspectral canopy sensing of paddy rice aboveground biomass at different growth stages[J]. Field Crops Research, 2014,155:42-55. doi: 10.1016/j.fcr.2013.09.023
[3]	张春桂,张星,陈敏艳,等.福建近岸海域悬浮泥沙浓度遥感定量监测研究[J].自然资源学报,2008,23(1):150-160. doi: 10.11849/zrzyxb.2008.01.017
	[ Zhang C G, Zhang X, Chen M Y, et al.Study on remote sensing quantitative model of suspended sediments in the coastal waters of Fujian[J]. Journal of Natural Resources, 2008,23(1):150-160. ] doi: 10.11849/zrzyxb.2008.01.017
[4]	An X F, Li M Z, Zheng L H, et al.Eliminating the interference of soil moisture and particle size on predicting soil total nitrogen content using a NIRS-based portable detector[J]. Computers and Electronics in Agriculture, 2015,112:47-53. doi: 10.1016/j.compag.2014.11.003
[5]	张娟娟. 土壤养分信息的光谱估测研究[D].南京:南京农业大学,2009.
	[ Zhang J J.Estimating soil nutrient information based on spectral analysis technology[D]. Nanjing: Nanjing Agricultural University, 2009. ]
[6]	Stenberg B, Rossel R A V, Mouazen A M, et al. Visible and near infrared spectroscopy in soil science[M]. Burlington: Academic Press, 2010.
[7]	王磊,白由路,卢艳丽,等.光谱数据变换对玉米氮素含量反演精度的影响[J].遥感技术与应用,2011,26(2):220-225.
	[ Wang L, Bai Y L, Lu Y L, et al.Effect on retrieval precision for corn N content by spectrum data transformation[J]. Remote Sensing Technology and Application, 2011,26(2):220-225. ]
[8]	Mashimbye Z E, Cho M A, Nell J P, et al.Model-based integrated methods for quantitative estimation of soil salinity from hyperspectral remote sensing data: a case study of selected South African soils[J]. Pedosphere, 2012,22(5):640-649.
[9]	贺军亮,蒋建军,周生路,等.土壤有机质含量的高光谱特性及其反演[J].中国农业科学,2007,40(3):638-643.
	[ He J L, Jiang J J, Zhou S L, et al.The hyperspectral characteristics and retrieval of soil organic matter content[J]. Scientia Agricultura Sinica, 2007,40(3):638-643. ]
[10]	Stevens A, Udelhoven T, Denis A, et al.Measuring soil organic carbon in croplands at regional scale using air borne imaging spectroscopy[J]. Geoderma, 2010,158(1):32-45. doi: 10.1016/j.geoderma.2009.11.032
[11]	Tian Y, Zhang J, Yao X, et al. Laboratory assessment of three quantitative methods for estimating the organic matter content of soils in China based on visible/near-infrared reflectance spectra[J]. Geoderma, 2013,202-203:161-170. doi: 10.1016/j.geoderma.2013.03.018
[12]	Kemper T, Sommer S.Estimate of heavy metal contamination in soils after a mining accident using reflectance spectroscopy[J]. Environmental Science & Technology, 2002,36(12):2742-2747. doi: 10.1021/es015747j pmid: 12099473
[13]	Nawar S, Buddenbaum H, Hill J, et al.Estimating the soil clay content and organic matter by means of different calibration methods of vis-NIR diffuse reflectance spectroscopy[J]. Soil & Tillage Research, 2016,155:510-522. doi: 10.1016/j.still.2015.07.021
[14]	Bayer A, Bachmann M, Müller A, et al.A comparison of feature-based MLR and PLS regression techniques for the prediction of three soil constituents in a degraded South African ecosystem[J]. Applied and Environmental Soil Science, 2012:971252. doi: 10.1155/2012/971252
[15]	Feng Y, Li X, Xu K, et al.Qualitative and simultaneous quantitative analysis of cimetidine polymorphs by ultraviolet-visible and shortwave near-infrared diffuse reflectance spectroscopy and multivariate calibration models[J]. Journal of Pharmaceutical and Biomedical Analysis, 2015,104:112-121.
[16]	Mohanty B, Gupta A, Das B S.Estimation of weathering indices using spectral reflectance over visible to mid-infrared region[J]. Geoderma, 2016,266:111-119. doi: 10.1016/j.geoderma.2015.11.030
[17]	Yu X, Liu Q, Wang Y, et al.Evaluation of MLSR and PLSR for estimating soil element contents using visible/near-infrared spectroscopy in apple orchards on the Jiaodong peninsula[J]. Catena, 2016,137:340-349. doi: 10.1016/j.catena.2015.09.024
[18]	鲁如坤. 土壤农业化学分析方法[M].北京:中国农业科技出版社,2000.
	[ Lu R K.Soil agricultural chemical analysis method[M]. Beijing: China's Agricultural Science and Technology Publishing, 2000. ]
[19]	Wold S.PLS for multivariate linear modeling[J]. Chemometric Methods in Molecular Design, 1995,2:195.
[20]	Chang C W, Laird D A.Near-infrared reflectance spectroscopic analysis of soil C and N[J]. Soil Science, 2002,167(2):110-116. doi: 10.1097/00010694-200202000-00003
[21]	周锦. 土壤中氮含量的测定分析[J].农业科技与信息,2008,15:40-41.
	[ Zhou J.Determination and analysis fornitrogen content in soil[J]. Information of Agricultural Science and Technology, 2008,15:40-41. ]
[22]	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 Y L, Bai Y L, Wang L, et al.Determination for total nitrogen content in black soil using hyperspectral data[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2010,26(1):256-261. ]
[23]	沈润平,丁国香,魏国栓,等.基于人工神经网络的土壤有机质含量高光谱反演[J].土壤学报,2009,46(3):391-397. doi: 10.3321/j.issn:0564-3929.2009.03.003
	[ Shen R P, Ding G X, Wei G S, et al.Retrieval of soil organic matter content from hyper-spectrum based on ANN[J]. Acta Pedologica Sinica, 2009,46(3):391-397. ] doi: 10.3321/j.issn:0564-3929.2009.03.003
[24]	杨爱霞,丁建丽.新疆艾比湖湿地土壤有机碳含量的光谱测定方法对比[J].农业工程学报,2015,31(18):162-168.
	[ Yang A X, Ding J L.Comparative assessment of two methods for estimation of soil organic carbon content by Vis-NIR spectra in Xinjiang Ebinur Lake Wetland[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2015,31(18):162-168. ]
[25]	Zornoza R, Guerrero C, Mataix-Solera J, et al.Near infrared spectroscopy for determination of various physical, chemical and biochemical properties in Mediterranean soils[J]. Soil Biology and Biochemistry, 2008,40(7):1923-1930. doi: 10.1016/j.soilbio.2008.04.003
[26]	Anderson K, Croft H.Remote sensing of soil surface properties[J]. Progress in Physical Geography, 2009,33(4):457-473. doi: 10.1177/0309133309346644
[27]	王镇浦,周国华,罗国安.偏最小二乘法(PLS)及其在分析化学中的应用[J].分析化学,1989,17(7):662-669. doi: 10.1007/BF02943117
	[ Wang Z P, Zhou G H, Luo G A.Partial least squares method (PLS) and its application in analytical chemistry[J]. Analytical Chemistry, 1989,17(7):662-669. ] doi: 10.1007/BF02943117
[28]	Wold S, Sjöström M, Eriksson L.PLS-regression: a basic tool of chemometrics[J]. Chemometrics and Intelligent Laboratory Systems, 2001,58(2):109-130. doi: 10.1016/S0169-7439(01)00155-1
[29]	Wang Q, Yi Q X, Bao A M, et al.Estimating chlorophyll density of cotton canopy by hyperspectral reflectance[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2012,28(15):125-132. doi: 10.3969/j.issn.1002-6819.2012.15.020
[30]	Brown D J, Shepherd, K D, Walsh M G, et al. Global soil characterization with VNIR diffuse reflectance spectroscopy[J]. Geoderma, 2006,132(3):273-290. doi: 10.1016/j.geoderma.2005.04.025
[31]	司海青,姚艳敏,王德营,等.含水率对土壤有机质含量高光谱估算的影响[J].农业工程学报,2015,31(9):114-120.
	[ Si H Q, Yao Y M, Wang D Y, et al.Hyperspectral prediction of soil organic matter contents under different soil moisture contents[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2015,31(9):114-120. ]

变量	模型	样本数	全距	最小值	最大值	平均值	标准差	方差
土壤氮素/（g·kg^-1）	校正集	99	0.50~5.59	0.50	5.59	1.30	1.05	1.10
土壤氮素/（g·kg^-1）	验证集	33	0.92~5.50	0.92	5.50	1.26	0.95	0.90

变换算法	校正模型参数		验证模型参数		最优因子个数
变换算法	R²	RMSE_C	R²	RMSE_V	最优因子个数
T₀	0.969192	0.000270	0.9843	0.000170	17
T₁	0.955778	0.000390	0.9825	0.000200	25
T₂	0.983095	0.000149	0.9927	0.000082	17
T₃	0.984178	0.000139	0.9888	0.000112	18
T₄	0.986729	0.000116	0.9933	0.000076	14
T₅	0.967169	0.000290	0.9789	0.000224	23
T₆	0.982890	0.000150	0.9897	0.000114	10
T₇	0.981017	0.000166	0.9858	0.000159	10
T₈	0.985022	0.000132	0.9853	0.000162	6
T₉	0.981608	0.000016	0.9865	0.000143	10
T₁₀	0.981803	0.000159	0.9886	0.000131	12
T₁₁	0.977755	0.000195	0.9709	0.000313	5
T₁₂	0.985058	0.000130	0.9848	0.000162	9
T₁₃	0.983277	0.000145	0.9865	0.000154	14
T₁₄	0.985637	0.000126	0.9903	0.000107	17
T₁₅	0.977619	0.000196	0.9900	0.000107	16