• Laser & Optoelectronics Progress
  • Vol. 60, Issue 17, 1730001 (2023)
Peng Xie, Zhenghai Wang*, Bei Xiao, and Yuxin Tian
Author Affiliations
  • School of Earth Sciences and Engineering, Sun Yat-sen University, Guangzhou 510275, Guangdong , China
  • show less
    DOI: 10.3788/LOP222037 Cite this Article Set citation alerts
    Peng Xie, Zhenghai Wang, Bei Xiao, Yuxin Tian. Hyperspectral Inversion of Soil Selenium Content Based on Seagull Algorithm Optimized Random Forest[J]. Laser & Optoelectronics Progress, 2023, 60(17): 1730001 Copy Citation Text show less
    Technology roadmap
    Fig. 1. Technology roadmap
    Spectral curves of soil samples. (a) Original spectra; (b) MSC-FD transformed spectra
    Fig. 2. Spectral curves of soil samples. (a) Original spectra; (b) MSC-FD transformed spectra
    SOA-RF Flowchart
    Fig. 3. SOA-RF Flowchart
    Correlation analysis. (a) Original spectra; (b) MSC-FD
    Fig. 4. Correlation analysis. (a) Original spectra; (b) MSC-FD
    sCARS algorithm feature band screening process
    Fig. 5. sCARS algorithm feature band screening process
    VISSA algorithm feature band screening process. (a) RMSECV values; (b) weight values
    Fig. 6. VISSA algorithm feature band screening process. (a) RMSECV values; (b) weight values
    Distribution of wavelength of features extracted by different algorithms
    Fig. 7. Distribution of wavelength of features extracted by different algorithms
    SOA-RF algorithm convergence curve
    Fig. 8. SOA-RF algorithm convergence curve
    SOA-RF model prediction results(RPD)
    Fig. 9. SOA-RF model prediction results(RPD)
    Scatter plot of measured and predicted values of the optimal model. (a) sCARS-PLSR; (b) IRIV-SVM; (c) sCARS-RF; (d) VCPA-GA-SOA-RF
    Fig. 10. Scatter plot of measured and predicted values of the optimal model. (a) sCARS-PLSR; (b) IRIV-SVM; (c) sCARS-RF; (d) VCPA-GA-SOA-RF
    SampleNumber of sampleMinimum value /(μg·g-1Maximum value /(μg·g-1Average /(μg·g-1Standard deviation
    Training set350.2001.1500.7260.256
    Prediction set140.2951.0250.5950.213
    Table 1. Descriptive statistics of soil selenium content in training and prediction sets
    Spectral transformTime /sNumber of variableTraining R2Training RMSEValidation R2Validation RMSE
    Full spectra20510.430.210.370.35
    sCARS<10.00100.520.170.490.19
    CARS<10.00140.520.190.400.16
    IRIV145.60130.650.150.490.18
    VISSA51.96650.490.190.480.16
    IVISSA27.11730.500.150.450.20
    VCPA-GA145.53420.540.150.510.19
    Table 2. Evaluation of accuracy of the PLSR model with different variable screenings
    Spectral transformTime /sNumber of variableTraining R2Training RMSEValidation R2Validation RMSE
    Full spectra20510.570.050.340.66
    sCARS<10.00100.590.110.570.14
    CARS<10.00140.580.130.550.12
    IRIV145.60130.660.130.550.11
    VISSA51.96650.770.130.530.13
    IVISSA27.11730.960.010.560.39
    VCPA-GA145.53420.560.110.540.14
    Table 3. Evaluation of accuracy of SVM models with different variable screenings
    Spectral transformTime /sNumber of variableTraining R2Training RMSEValidation R2Validation RMSE
    Full spectra20510.850.190.470.18
    sCARS<10.00100.820.090.660.13
    CARS<10.00140.620.160.490.17
    IRIV145.60130.860.060.680.15
    VISSA51.96650.860.050.650.16
    IVISSA27.11730.810.060.640.15
    VCPA-GA145.53420.690.110.590.19
    Table 4. Evaluation of accuracy of the RF model with different variable screenings
    Spectral transformTime /sNumber of variableTraining R2Training RMSEValidation R2Validation RMSE
    Full spectra20510.850.160.750.15
    sCARS<10.00100.930.070.900.06
    CARS<10.00140.910.080.800.11
    IRIV145.60130.860.100.810.11
    VISSA51.96650.940.070.840.10
    IVISSA27.11730.940.060.830.09
    VCPA-GA145.53420.940.070.920.08
    Table 5. Evaluation of accuracy of the SOA-RF model with different variable screenings
    Peng Xie, Zhenghai Wang, Bei Xiao, Yuxin Tian. Hyperspectral Inversion of Soil Selenium Content Based on Seagull Algorithm Optimized Random Forest[J]. Laser & Optoelectronics Progress, 2023, 60(17): 1730001
    Download Citation