Side-by-side Chinese-English

消除水分因素影响的野外原状土壤盐分高光谱建模估测

陈红艳1 赵庚星1 李玉环1 李华2 盖岳峰3

(1.土肥资源高效利用国家工程实验室/山东农业大学资源与环境学院, 泰安 271018)
(2.山东菏泽水利工程总公司, 菏泽 274000)
(3.山东颐通土地房地产评估测绘有限公司, 济南 250000)
【知识点链接】偏最小二乘回归

【摘要】土壤水分被确定为土壤属性 (有机碳、盐分等) 光谱预测准确性下降的一个主要原因, 该文通过两种方法的对比旨在探索去除土壤水分影响、提高盐分高光谱定量估测精度的方法和技术路线。首先以山东省东营市垦利区为研究区, 采用地物光谱仪测定了96个样本的野外原状土和室内风干土光谱, 并进行一阶导数变换;接着, 对比分析盐分光谱特征及水分的影响;然后分别采用外部参数正交化 (external parameter orthogonalization, EPO) 和非负矩阵分解 (non-negative matrix factorizing, NMF) 方法校正和融合野外原状土光谱, 去除土壤水分因素的影响, 形成野外原状土光谱的校正和融合光谱;最后基于野外原状土光谱、校正和融合光谱, 分别采用多元逐步线性回归 (multiple step linear regression, MSLR) 和偏最小二乘回归 (partial least squares regression, PLSR) 构建土壤盐分含量的估测模型, 并进行验证和比较, 分析预测精度变化。结果显示:土壤水分对野外原状土光谱及盐分光谱特征影响较大, 需要研究去除;EPO和NMF均可提高土壤盐分野外原位光谱估测精度, 比较而言, NMF效果更为显著;EPO结合PLSR或NMF结合MSLR可作为去除水分影响的土壤盐分校准模型的技术路线。

【关键词】 土壤盐分;土壤水分;光谱分析;外部参数正交化;非负矩阵分解;

【DOI】

【基金资助】 国家自然科学基金项目 (41401239, 41671346) ; 国家科技支撑计划 (2015BAD23B02) ; 山东农业大学“双一流”奖补资金资助 (SYL2017XTTD02) ; 山东省重点研发计划 (2017CXGC0306) ;

Modeling and estimation of field undisturbed soil salt based on hyperspectra under removal of moisture factor

【Abstract】Soil moisture (SM) is one of the main reasons for the decline of predictive accuracy of soil attributes (organic carbon and salt, etc.) in using the spectrum analysis method. By comparing the two methods of external parameter orthogonization (EPO) and non-negative matrix factorizing (NMF), we explored a method and technology route of removing the effect of SM and improving the estimation precision of soil salt content (SSC) based on hyperspectra. Firstly, we used Kenli district of Dongying City in Shandong Province as the research area, and took 96 soil samples in the fields. The samples’ hyperspectra in situ and indoor after air-drying were measured respectively by spectra radiometer, and then transformed to the first deviation. The contents of soil salt and moisture were measured in laboratory. Then, the spectral characteristics of SSC and the effect of SM on it were analyzed by comparison. Next, the EPO and NMF were respectively used to correct and fuse the soil spectra in-situ (Situ-spectra), and to remove the SM effect and form the EPO correction spectra (EPO-spectra) and the NMF fusion spectra (NMF-spectra) of the Situ-spectra. Finally, the estimation models of SSC were built respectively by the multiple step linear regression (MSLR) and the partial least squares regression (PLSR) based on the Situ-spectra, EPO-spectra and NMF-spectra, and were verified and compared to analyze the changes of the SSC prediction precision. The results indicated that the SSC was high, the SSC gradient was obvious, and the dispersion degree of SSC was high. However, the SM content was about 30 folds of the SSC in the study area. The correlation between soil salt and spectra was better at the wavelength ranges of 1 440–1 660 nm, 1 830–1 860 nm, 1 960–2 110 nm. The SM had great effects on the Situ-spectra and SSC spectral characteristics. Therefore, it is necessary to remove SM impact. The EPO method could reduce the correlation between spectra and SM in most spectral regions, and at the same time weaken the correlation between spectra and soil salt in local wavelengths. In comparison, the NMF method could effectively reduce the correlation between spectra and SM, and increase the correlation between spectra and soil salinity. Both the EPO and NMF could improve the accuracy of the SSC estimation based on Situ-spectra. After adoption of EPO, the validation coefficient of determination (R2) was increased by 0.08–0.09, and the relative prediction deviation (RPD) was increased by 0.08–0.69. At the same time, after adoption of NMF, the validation R2 was increased by 0.27–0.38, reaching above 0.80, and the RPD was increased by 1.04–1.06, reaching above of 2.37. Thus, the result of NMF was more significant than that of EPO for the removal of the SM effect. The method of EPO combined with PLSR or NMF combined with MSLR could be used as the technical route of removing the SM effect and building the SSC correction model. The results can effectively promote the quantitative remote sensing extraction and real-time and in-situ monitoring of the saline soil information.

【Keywords】 soil salt; soil moisture; spetrum analysis; external parameter orthogonalization; non-negative matrix factorizing;

【DOI】

【Funds】 National Natural Science Foundation of China (41401239, 41671346); National Key Technology Research and Development Program of the Ministry of Science and Technology of China (2015BAD23B02); “Double First-Class” Award Funding of Shandong Agricultural University (SYL2017XTTD02); Key Research and Development Program in Shandong Province (2017CXGC0306);

Download this article
    References

    [1] Shi Z, Guo Y, Jin X, et al. Advancement in study on proximal soil sensing [J]. Acta Pedologica Sinica, 2011, 48 (6): 1274–1281 (in Chinese with English abstract).

    [2] Ignacio Melendez-Pastor, Encarni I Hernández, Jose Navarro-Pedreño, et al. Mapping soil salinization of agricultural coastal areas in Southeast Spain [M/OL] //Dr. Boris Escalante (Ed.), Remote Sensing-Applications, 2012: 117–140. http://www.intechopen. com/books/-remotesensing-applications/mapping-soil-salinization-of-agricultural-coastal-areas-in-southeast-spain

    [3] Farifteh J, Van Der Meer F, Van Der Meijde M, et al. Spectral characteristics of salt-affected soils: A laboratory experiment [J]. Geoderma, 2008, 145: 196–206.

    [4] Nawar S, Buddenbaum H, Hill J. Estimation of soil salinity using three quantitative methods based on visible and near-infrared reflectance spectroscopy: A case study from Egypt [J]. Arabian Journal of Geosciences, 2015, 8 (7): 5127–5140.

    [5] Dai XJ, Peng J, Zhang YL, et al. Prediction on soil salt content based on spectral classification [J]. Acta Pedologica Sinica, 2016, 53 (4): 909–918 (in Chinese with English abstract).

    [6] Zhang XL, Zhang F, Zhang HW, et al. Optimization of soil salt inversion model based on spectral transformation from hyperspectral index [J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2018, 34 (1): 110–117 (in Chinese with English abstract).

    [7] Iii J B R, Smith D B. The potential of mid-and near-infrared diffuse reflectance spectroscopy for determining major- and trace-element concentrations in soils from a geochemical survey of North America [J]. Applied Geochemistry, 2009, 24 (8): 1472–1481.

    [8] Bambangh K, Hedley C B, Hedley M J, et al. The use of diffuse reflectance spectroscopy for in situ carbon and nitrogen analysis of pastoral soils [J]. Australian Journal of Soil Research, 2008, 46 (6): 623–635.

    [9] Ji W, Rossel R A V, Shi Z. Accounting for the effect of water and the environment on proximally sensed vis-NIR soil spectra and their calibrations [J]. European Journal of Soil Science, 2015, 66 (3): 555–565.

    [10] Minasny B, McBratney A B, Bellon-Maurel V, et al. Removing the effect of soil moisture from NIR diffuse reflectance spectra for the prediction of soil organic carbon [J]. Geoderma, 2011, 167/168: 118–124

    [11] Wang Q, Li P H, Chen X. Modeling salinity effect on soil reflectance under various moisture conditions and its inverse application: A laboratory experiment [J]. Geoderma, 2012, 170: 103–111.

    [12] Nocita M, Stevens A, Noon C, et al. Prediction of soil organic carbon for different levels of soil moisture using Vis-NIR spectroscopy [J]. Geoderma, 2013, 199: 37–42.

    [13] Rienzi E A, Mijatovic B, Mueller T G, et al. Prediction of soil organic carbon under varying moisture levels using reflectance spectroscopy [J]. Soil Science Society of America Journal, 2014, 78 (3): 958–967.

    [14] Castaldi F, Palombo A, Pascucci S, et al. Reducing the influence of soil moisture on the estimation of clay from hyperspectral data: A case study using simulated PRISMA Data [J]. Remote Sensing, 2015, 7 (11): 15561–15582.

    [15] Kanzari S, Hachicha M, Bouhlila R, et al. Characterization and modeling of water movement and salts transfer in a semi-arid region of Tunisia (Bou Hajla, Kairouan)—Salinization risk of soils and aquifers [J]. Computers & Electronics in Agriculture, 2012, 86 (s4–6): 34–42.

    [16] Liu Y, Pan X Z, Shi R J, et al. Predicting soil salt content over partially vegetated surfaces using non-negative matrix factorization [J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2016, 8 (11): 5305–5316.

    [17] Ji W, Viscarra Rossel R A, Shi Z. Improved estimates of organic carbon using proximally sensed Vis-NIR spectra corrected by piecewise direct standardization [J]. European Journal of Soil Science, 2015, 66 (4): 670–678.

    [18] Ackerson J P, Dematte J A, Morgan C L. Predicting clay content on field-moist intact tropical soils using a dried, ground Vis-NIR library with external parameter orthogonalization [J]. Geoderma, 2015, s259–260: 196–204.

    [19] Wijewardane N K, Ge Y, Morgan C L S. Moisture insensitive prediction of soil properties from reflectance spectra based on external parameter orthogonalization [J]. Geoderma, 2016, 267: 92–101.

    [20] Roudier P, Hedley C B, Lobsey C R, et al. Evaluation of two methods to eliminate the effect of water from soil vis-NIR spectra for predictions of organic carbon [J]. Geoderma, 2017, 296: 98–107.

    [21] Lee D D, Seung H S. Learning the parts of objects by non-negative matrix factorization [J]. Nature, 1999, 401 (6755): 788–791.

    [22] Chen HY, Zhao GX, Chen JC, et al. Remote sensing inversion of saline soil salinity based on modified vegetation index in estuary area of Yellow River [J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2015, 31 (5): 107–114 (in Chinese with English abstract).

    [23] Chen B, Zou XY, Zhu WJ. Eliminating outlier samples in near-infrared model by method of PCA–Mahalanobis distance [J]. Journal of Jiangsu University: Natural Science Edition, 2008, 29 (4): 277–279 (in Chinese with English abstract).

    [24] Peng J, Wang JQ, Xiang HY, et al. Comparative study on hyperspectral inversion accuracy of soil salt content and electrical conductivity [J]. Spectroscopy and Spectral Analysis, 2014, 34 (2): 510–514 (in Chinese with English abstract).

    [25] Ge Y, Morgan C L, Ackerson J P. Vis-NIR spectra of dried ground soils predict properties of soils scanned moist and intact [J]. Geoderma, 2014, 221: 61–69.

    [26] Huang M, Zhu QB. Feature extraction of hyperspectral scattering image foe apple mealiness based on singular value decomposition [J]. Spectroscopy and Spectral Analysis, 2011, 31 (3): 767–770 (in Chinese with English abstract).

    [27] Chen HY, Zhao GX, Zhang XH, et al. Improving estimation precision of soil organic matter content by removing effect of soil moisture from hyperspectra [J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2014, 30 (8): 91–100 (in Chinese with English abstract).

    [28] Maruyama R, Maeda K, Moroda H, et al. Detecting cells using non-negative matrix factorization on calcium Imaging data [J]. Neural Netw, 2014, 55 (55c): 11.

    [29] Galeano J, Jolivot R, Marzani F, et al. Unmixing of human skin optical reflectance maps by non-negative matrix factorization algorithm [J]. Biomedical Signal Processing & Control, 2013, 8 (2): 169–175.

    [30] Lin C J. On the convergence of multiplicative update algorithms for non–negative matrix factorization [J]. IEEE Transactions on Neural Networks, 2007, 18 (6): 1589–1596.

    [31] Jiang ZL, Yang YS, Sha JM. Study on GWRmodel applied for hyperspectral prediction of soil chromium in Fuzhou City [J]. Acta Ecologica Sinica, 2017, 37 (23): 8117–8127 (in Chinese with English abstract).

    [32] Wang JZ, Tashpolat·Tiyip, Ding JL, et al. Estimation of desert soil organic carbon content based on hyperspectral data preprocessing with fractional differential [J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2016, 32 (21): 161–169 (in Chinese with English abstract).

    [33] Sakhaii P, Bermel W. Improving the sensitivity of conventional spin echospectra by preservation of initial signal-to-noise ratio [J]. Journal of Magnetic Resonance, 2014, 242 (3): 220–223.

    [34] Liu Y, Pan XZ, Wang CK, et al. Predicting soil salinity based on spectral system under wet soil condition [J]. Spectroscopy and Spectral Analysis, 2013, 33 (10): 2771–2776 (in Chinese with English abstract).

    [35] Li Y L, Liu Y, Wu S W, et al. Hyper-spectral estimation of wheat biomass after alleviating of soil effect on spectra by non-negative matrix factorization [J]. European Journal of Agronomy, 2017, 84: 58–66.

This Article

ISSN:1002-6819

CN: 11-2047/S

Vol 34, No. 12, Pages 119-125

June 2018

Downloads:0

Share
Article Outline

Knowledge

Abstract

  • 0 Introduction
  • 1 Materials and methods
  • 2 Results and analysis
  • 3 Discussion
  • 4 Conclusions
  • References