Hyperspectral reflectance models for soil salt content by filtering methods and waveband selection

Open access

Abstract

For improving the understanding of interactions between hyperspectral reflectance and soil salinity, in situ hyperspectral inversion of soil salt content at a depth of 0-10 cm was conducted in Hetao Irrigation District, Inner Mongolia, China. Six filtering methods were used to preprocess soil reflectance data, and waveband selection combined by VIP (variable importance in projection) and b-coefficients (regression coefficients of model) was also applied to simplify model. Then statistical methods of partial least square regression (PLS) and orthogonal projection to latent structures (OPLS) were processed to establish the inversion models. Our findings indicate that the selected sensitive wavebands for the 6 filtering methods are different, among which the multiplicative signal correction (MSC) and standard normal variate methods (SNV) have some similar sensitive wavebands with unfiltered data. Derivatives (DF1 and DF2) could characterize sensitive wavebands along the scale of VNIR (350-1100 nm), especially the second derivative (DF2). The sensitive wavebands for continuum-removed reflectance method (CR) have protruded many narrow absorption features. For orthogonal signal correction method (OSC), the selected wavebands are centralized in the range of 565-1013 nm. The calibration and evaluation processes have demonstrated the second order derivate filtering method (DF2) combined with waveband selection is superior to other processes, for it has high R2 (larger than 0.7) both in PLS and OPLS models for calibration and evaluation, by choosing only 156 wavebands from the whole 700 wavebands. Meanwhile, OPLS method was considered to be more suitable for the analyzing than PLS in most of our situations.

[1] Peragón JM, Delgado A, Díaz JaR, Pérez-Latorre FJ. A GIS-based decision tool for reducing salinization risks in olive orchards. Agr Water Manage. 2016;166:33-41. DOI:10.1016/j.agwat.2015.12.005.

[2] Benini L, Antonellini M, Laghi M, Mollema P. Assessment of water resources availability and groundwater salinization in future climate and land use change scenarios: A case study from a coastal drainage basin in Italy. Water Resour Manag. 2015;30(2):731-745. DOI: 10.1007/s11269-015-1187-4.

[3] Song D, Liu B, Li X, Chen S, Li L, Ma M, et al. Hyperspectral data spectrum and texture band selection based on the subspace-rough set method. Int J Remote Sens. 2015;36(8):2113-2128. DOI: 10.1080/01431161.2015.1034892.

[4] Kattenborn T, Maack J, Faßnacht F, Enßle F, Ermert J, Koch B. Mapping forest biomass from space-Fusion of hyperspectral EO1-hyperion data and Tandem-X and WorldView-2 canopy height models. International J Appl Earth Observation Geoinform. 2015;35:359-367. DOI: 10.1016/j.jag.2014.10.008.

[5] Dehaan R, Taylor G. Image-derived spectral endmembers as indicators of salinisation. Int J Remote Sens. 2003;24,(4):775-794. DOI: 10.1080/01431160110107635.

[6] Csillag F, Pásztor L, Biehl LL. Spectral band selection for the characterization of salinity status of soils. Remote Sens Environ. 1993,43(3):231-242. DOI: 10.1016/0034-4257(93)90068-9.

[7] Pang G, Wang T, Liao J, Li S. Quantitative model based on field-derived spectral characteristics to estimate soil salinity in Minqin County, China. Soil Sci Soc Am J. 2014;78,(2):546-555. DOI: 10.2136/sssaj2013.06.0241.

[8] Hunt GR. Spectral signatures of particulate minerals in the visible and near infrared. Geophysics. 1977;42,(3):501-513. DOI: 10.1190/1.1440721.

[9] Hick P, Russell W. Some spectral considerations for remote sensing of soil salinity. Soil Res. 1990;28,(3):417-431. DOI:10.1071/SR9900417.

[10] Hirschfeld T. Salinity determination using NIRA. Appl Spectrosc. 1985;39(4):740-741. DOI: 10.1366/0003702854250293.

[11] Tsai F, Philpot W. Derivative analysis of hyperspectral data. Remote Sens Environ. 1998;66(1):41-51. DOI: 10.1016/S0034-4257(98)00032-7.

[12] Andersson M. A comparison of nine PLS1 algorithms. J Chemometr. 2009;23(10):518-529. DOI: 10.1002/cem.1248.

[13] Dumarey M, Goodwin DJ, Davison C. Multivariate modelling to study the effect of the manufacturing process on the complete tablet dissolution profile. Int J Pharm. 2015;486(1):112-120. DOI: 10.1016/j.ijpharm.2015.03.040.

[14] Gabrielsson J, Jonsson H, Airiau C, Schmidt B, Escott R, Trygg J. OPLS methodology for analysis of pre-processing effects on spectroscopic data. Chemometr Intell Lab. 2006;84(1):153-158. DOI: 10.1016/j.chemolab.2006.03.013.

[15] Trygg J, Wold S. Orthogonal projections to latent structures (O-PLS). J Chemometr. 2002;16(3):119-128. DOI: 10.1002/cem.695.

[16] Gosselin R, Rodrigue D, Duchesne C. A Bootstrap-VIP approach for selecting wavelength intervals in spectral imaging applications. Chemometr Intell Lab. 2010;100(1):12-21. DOI: 10.1016/j.chemolab.2009.09.005.

[17] Haaland DM, Thomas EV. Partial least-squares methods for spectral analyses. 2. Application to simulated and glass spectral data. Anal Chem. 1988;60(11):1202-1208. DOI: 10.1021/ac00162a021.

[18] Lilliefors HW. On the Kolmogorov-Smirnov test for normality with mean and variance unknown. J Am Stat Assoc. 1967;62(318):399-402. DOI: 10.1080/01621459.1967.10482916.

[19] Martens H, Nielsen JP, Engelsen SB. Light scattering and light absorbance separated by extended multiplicative signal correction. Application to near-infrared transmission analysis of powder mixtures. Anal Chem. 2003;75(3):394-404. DOI: 10.1021/ac020194w.

[20] Dhanoa M, Lister S, Sanderson R, Barnes R. The link between multiplicative scatter correction (MSC) and standard normal variate (SNV) transformations of NIR spectra. J Near Infrared Spectrosc. 1994;2(1):42-47. DOI: 10.1255/jnirs.30.

[21] Inoue Y, Sakaiya E, Zhu Y, Takahashi W. Diagnostic mapping of canopy nitrogen content in rice based on hyperspectral measurements. Remote Sens Environ. 2012;126:210-221. DOI: 10.1016/j.rse.2012.08.026.

[22] Vašát R, Kodešová R, Borůvka L, Klement A, Jakšík O, Gholizadeh A. Consideration of peak parameters derived from continuum-removed spectra to predict extractable nutrients in soils with visible and near-infrared diffuse reflectance spectroscopy (VNIR-DRS). Geoderma. 2014;232:208-218. DOI: 10.1016/j.geoderma.2014.05.012.

[23] Wang G, Yin S. Quality-related fault detection approach based on orthogonal signal correction and modified PLS. IEEE Trans Industr Informatics. 2015;11(2):398-405. DOI: 10.1109/TII.2015.2396853.

[24] Eastment H, Krzanowski W. Cross-validatory choice of the number of components from a principal component analysis. Technometrics. 1982;24(1):73-77. DOI: 10.1080/00401706.1982.10487712.

[25] Gomez C, Lagacherie P, Coulouma G. Continuum removal versus PLSR method for clay and calcium carbonate content estimation from laboratory and airborne hyperspectral measurements. Geoderma. 2008;148(2):141-148. DOI: 10.1016/j.geoderma.2008.09.016.

[26] Whiting ML, Li L, Ustin SL. Predicting water content using Gaussian model on soil spectra. Remote Sens Environ. 2004;89(4):535-552. DOI: 10.1016/j.rse.2003.11.009.

[27] Elmasry G, Wang N, Vigneault C, Qiao J, Elsayed A. Early detection of apple bruises on different background colors using hyperspectral imaging. LWT-Food Science and Technology. 2008;41(2):337-345. DOI: 10.1016/j.lwt.2007.02.022.

[28] Griffin JL. Metabonomics: NMR spectroscopy and pattern recognition analysis of body fluids and tissues for characterisation of xenobiotic toxicity and disease diagnosis. Curr Opin Chem Biol. 2003;7(5):648-654. DOI: 10.1016/j.cbpa.2003.08.008.

[29] Lin W-S, Yang C-M, Kuo B-J. Classifying cultivars of rice (Oryza sativa L.) based on corrected canopy reflectance spectra data using the orthogonal projections to latent structures (O-PLS) method. Chemometr Intell Lab. 2012;115:25-36. DOI: 10.1016/j.chemolab.2012.04.005.

[30] Tapp HS, Kemsley EK. Notes on the practical utility of OPLS. TrAC Trends Anal Chem. 2009;28(11):1322-1327. DOI: 10.1016/j.trac.2009.08.006.

[31] Cho MA, Skidmore A, Corsi F, Van Wieren SE, Sobhan I. Estimation of green grass/herb biomass from airborne hyperspectral imagery using spectral indices and partial least squares regression. Int J Appl Earth Observ Geoinformat. 2007;9(4):414-424. DOI: 10.1016/j.jag.2007.02.001.

Ecological Chemistry and Engineering S

The Journal of Society of Ecological Chemistry and Engineering

Journal Information


IMPACT FACTOR 2017: 0.7
5-year IMPACT FACTOR: 0.815

CiteScore 2017: 0.79

SCImago Journal Rank (SJR) 2017: 0.227
Source Normalized Impact per Paper (SNIP) 2017: 0.535

Cited By

Metrics

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 167 165 12
PDF Downloads 70 70 5