• Spectroscopy and Spectral Analysis
  • Vol. 42, Issue 6, 1704 (2022)
Wen-qiang SHI1、*, Xiu-ying XU1、1; *;, Wei ZHANG1、1;, Ping ZHANG2、2;, Hai-tian SUN1、1; 3;, and Jun HU1、1;
Author Affiliations
  • 11. College of Engineering, Heilongjiang Bayi Agricultural University, Daqing 163319, China
  • 22. College of Science, Heilongjiang Bayi Agricultural University, Daqing 163319, China
  • show less
    DOI: 10.3964/j.issn.1000-0593(2022)06-1704-07 Cite this Article
    Wen-qiang SHI, Xiu-ying XU, Wei ZHANG, Ping ZHANG, Hai-tian SUN, Jun HU. Prediction Model of Soil Moisture Content in Northern Cold Region Based on Near-Infrared Spectroscopy[J]. Spectroscopy and Spectral Analysis, 2022, 42(6): 1704 Copy Citation Text show less
    Original near infrared spectra of soil
    Fig. 1. Original near infrared spectra of soil
    Preprocessed near infrared spectra of soil(a): Preprocessed by MSC; (b): Preprocessed by SNV;(c): Preprocessed by first derivative; (d): Preprocessed by second derivative;(e): Preprocessed by smoothing
    Fig. 2. Preprocessed near infrared spectra of soil
    (a): Preprocessed by MSC; (b): Preprocessed by SNV;(c): Preprocessed by first derivative; (d): Preprocessed by second derivative;(e): Preprocessed by smoothing
    The best training effects based on different algorithms(a): S_G-BP model; (b): SNV-SVM model; (c): S_G-GP model
    Fig. 3. The best training effects based on different algorithms
    (a): S_G-BP model; (b): SNV-SVM model; (c): S_G-GP model
    The best prediction fitting effects based on different algorithms(a): S_G-SVM model; (b): BP model; (c): MSC-GP model
    Fig. 4. The best prediction fitting effects based on different algorithms
    (a): S_G-SVM model; (b): BP model; (c): MSC-GP model
    Comparison of prediction results(a): Based on original data; (b): Based on MSC preprocessing;(c): Based on SNV preprocessing; (d): Based on first derivative preprocessing;(e): Based on the preprocessing of second derivative; (f): Based on smoothing preprocessing
    Fig. 5. Comparison of prediction results
    (a): Based on original data; (b): Based on MSC preprocessing;(c): Based on SNV preprocessing; (d): Based on first derivative preprocessing;(e): Based on the preprocessing of second derivative; (f): Based on smoothing preprocessing
    样本集合样本数最小值最大值平均值标准差
    建模集6915.00050.00032.60911.408
    测试集3515.00050.00032.28611.548
    总样本集10415.00050.00032.50011.456
    Table 1. Basic information of soil moisture
    模型建模集测试集
    Rc2RMSECRp2RMSEP
    SVM0.989 81.166 20.991 10.786 8
    MSC-SVM0.990 01.143 60.990 30.810 4
    SNV-SVM0.991 11.081 50.988 10.877 1
    D1-SVM0.988 21.221 20.991 80.745 8
    D2-SVM0.988 41.218 20.992 30.737 6
    S_G-SVM0.989 51.202 70.992 10.736 9
    BP0.958 72.356 00.969 71.544 8
    MSC-BP0.958 22.372 90.962 11.627 1
    SNV-BP0.958 92.429 00.959 81.713 6
    D1-BP0.950 92.563 00.964 91.546 1
    D2-BP0.958 12.425 50.955 51.787 5
    S_G-BP0.960 92.379 70.957 01.682 0
    GP0.920 73.300 10.925 22.219 2
    MSC-GP0.907 33.509 50.934 82.148 6
    SNV-GP0.912 23.433 80.911 32.512 0
    D1-GP0.919 13.348 60.912 72.630 0
    D2-GP0.907 13.314 40.921 92.503 2
    S_G-GP0.928 03.258 10.909 32.782 9
    Table 2. Comparison of different model parameters
    Wen-qiang SHI, Xiu-ying XU, Wei ZHANG, Ping ZHANG, Hai-tian SUN, Jun HU. Prediction Model of Soil Moisture Content in Northern Cold Region Based on Near-Infrared Spectroscopy[J]. Spectroscopy and Spectral Analysis, 2022, 42(6): 1704
    Download Citation