Machine learning based prediction of soil total nitrogen, organic carbon and moisture content by using VIS-NIR spectroscopy

被引:333
|
作者
Morellos, Antonios [1 ]
Pantazi, Xanthoula-Eirini [1 ]
Moshou, Dimitrios [1 ]
Alexandridis, Thomas [3 ]
Whetton, Rebecca [2 ]
Tziotzios, Georgios [1 ]
Wiebensohn, Jens [4 ]
Bill, Ralf [4 ]
Mouazen, Abdul M. [2 ]
机构
[1] Aristotle Univ Thessaloniki, Agr Engn Lab, Fac Agr, Univ Box 275, Thessaloniki 54124, Greece
[2] Cranfield Univ, Cranfield Soil & AgriFood Inst, Cranfield MK43 0AL, Beds, England
[3] Aristotle Univ Thessaloniki, Lab Remote Sensing & GIS, Fac Agr, Univ Box 259, Thessaloniki 54124, Greece
[4] Univ Rostock, Professorship Geodesy & Geoinformat, Fac Agr & Environm Sci, D-18055 Rostock, Germany
关键词
VIS-NIR spectroscopy; Data mining; Chemometrics; Soil properties; NEAR-INFRARED SPECTROSCOPY; PRINCIPAL COMPONENT ANALYSIS; PARTIAL LEAST-SQUARES; REFLECTANCE SPECTROSCOPY; NEURAL-NETWORK; ONLINE; PH;
D O I
10.1016/j.biosystemseng.2016.04.018
中图分类号
S2 [农业工程];
学科分类号
0828 ;
摘要
It is widely known that the visible and near infrared (VIS-NIR) spectroscopy has the potential of estimating soil total nitrogen (TN), organic carbon (OC) and moisture content (MC) due to the direct spectral responses these properties have in the near infrared (NIR) region. However, improving the prediction accuracy requires advanced modelling techniques, particularly when measurement is planned for fresh (wet and un-processed) soil samples. The aim of this work is to compare the predictive performance of two linear multivariate and two machine learning methods for TN, OC and MC. The two multivariate methods investigated included principal component regression (PCR) and partial least squares regression (PLSR), whereas the machine learning methods included least squares support vector machines (LS-SVM), and Cubist. A mobile, fibre type, VIS-NIR spectrophotometer was utilised to collect soil spectra (305-2200 nm) in diffuse reflectance mode from 140 wet soil samples collected from one field in Germany. The results indicate that machine learning methods are capable of tackling non-linear problems in the dataset. LS-SVMs and the Cubist method out-performed the linear multivariate methods for the prediction of all three soil properties studied. LS-SVM provided the best prediction for MC (root mean square error of prediction (RMSEP) = 0.457% and residual prediction deviation (RPD) = 2.24) and OC (RMSEP = 0.062% and RPD = 2.20), whereas the Cubist method provided the best prediction for TN (RMSEP = 0.071 and RPD = 1.96). (C) 2016 IAgrE. Published by Elsevier Ltd. All rights reserved.
引用
收藏
页码:104 / 116
页数:13
相关论文
共 50 条
  • [31] Feature Variable Selection Based on VIS-NIR Spectra and Soil Moisture Content Prediction Model Construction
    Zhou, Nan
    Hong, Jin
    Song, Bo
    Wu, Shichao
    Wei, Yichen
    Wang, Tao
    JOURNAL OF SPECTROSCOPY, 2024, 2024
  • [32] VIS-NIR spectroscopy and environmental factors coupled with PLSR models to predict soil organic carbon and nitrogen
    Zhu, Jingrong
    Jin, Yihua
    Zhu, Weihong
    Lee, Dong Kun
    INTERNATIONAL SOIL AND WATER CONSERVATION RESEARCH, 2024, 12 (04) : 844 - 854
  • [33] Vis-NIR Spectroscopy for Soil Organic Carbon Assessment: A Meta-Analysis
    Chinilin, A. V.
    Vindeker, G. V.
    Savin, I. Yu.
    EURASIAN SOIL SCIENCE, 2023, 56 (11) : 1605 - 1617
  • [34] Sampling for field measurement of soil carbon using Vis-NIR spectroscopy
    Singh, Kanika
    Minasny, Budiman
    McBratney, Alex B.
    Short, Michael G.
    Namazi, Fatemeh
    DIGITAL SOIL ASSESSMENTS AND BEYOND, 2012, : 415 - 420
  • [35] Optimal sample selection for measurement of soil organic carbon using online vis-NIR spectroscopy
    Nawar, Said
    Mouazen, Abdul M.
    COMPUTERS AND ELECTRONICS IN AGRICULTURE, 2018, 151 : 469 - 477
  • [36] Vis-NIR Spectroscopy for Soil Organic Carbon Assessment: A Meta-Analysis
    A. V. Chinilin
    G. V. Vindeker
    I. Yu. Savin
    Eurasian Soil Science, 2023, 56 : 1605 - 1617
  • [37] PREDICTION OF ORGANIC CARBON AND CALCIUM CARBONATES IN AGRICULTURAL SOILS WITH VIS-NIR SPECTROSCOPY
    Milos, B.
    Bensa, A.
    POLJOPRIVREDA, 2018, 24 (01): : 45 - 51
  • [38] Evaluation of spectral data based soil organic carbon content estimation models in VIS-NIR
    Nagy, Attila
    Szabo, Andrea
    Escobar, Diana Quintin
    Tamas, Janos
    SOIL SCIENCE ANNUAL, 2024, 75 (01)
  • [39] Estimating forest soil organic carbon content using vis-NIR spectroscopy: Implications for large-scale soil carbon spectroscopic assessment
    Liu, Shangshi
    Shen, Haihua
    Chen, Songchao
    Zhao, Xia
    Biswas, Asim
    Jia, Xiaolin
    Shi, Zhou
    Fang, Jingyun
    GEODERMA, 2019, 348 : 37 - 44
  • [40] Removing the moisture effect on predicting soil organic matter using vis-NIR spectroscopy with external parameter orthogonalization
    Yang, Meihua
    Chen, Songchao
    Xu, Dongyun
    Zhao, Xiaomin
    Shi, Zhou
    Qian, Haiyan
    Zhang, Zhi
    GEODERMA REGIONAL, 2024, 37