Laser & Optoelectronics Progress, Volume. 62, Issue 17, 1739023(2025)

Spectral Feature Equilibrium for Enhanced Lead Detection in Kelp via LIBS-LIF (Invited)

Lingling Pi1, Shengqun Shi2, Weihua Huang2, Zhiyong Ouyang2, Junfei Nie3, Jinling Xiao1, and Lianbo Guo2、*
Author Affiliations
  • 1School of Optical and Electronic Information, Huazhong University of Science and Technology, Wuhan 430074, Hubei , China
  • 2Wuhan National Laboratory for Optoelectronics, Huazhong University of Science and Technology, Wuhan 430074, Hubei , China
  • 3School of Electrical Engineering, Shaoyang University, Shaoyang 422000, Hunan , China
  • show less
    Figures & Tables(9)
    Schematic diagrams of experimental setup and sample preparation process. (a) Schematic diagram of LIBS-LIF system setup;
    Spectra of kelp samples. (a) LIBS spectra; (b) LIBS-LIF spectra
    Prediction results based on LIBS-LIF univariate quantitative analysis and bispectral fusion multivariate quantitative analysis.
    Feature importance analysis and spectra of LIBS-LIF and LIBS after fusion. (a) Schematic diagram of feature importance; (b) schematic diagram of fusion spectra
    Schematic diagram of feature equilibrium processing process. (a) Schematic diagram of the original spectral fusion; (b) schematic diagram of the spectra before and after the interpolation of the LIBS-LIF spectrum and the selection of the peak fragment for LIBS spectrum; (c) SVR multiple regression importance analysis result of characteristic equilibrium spectra
    Performance analysis of feature equilibrium method in SVR regression
    Comparison of prediction performances for different regression models. (a) RMSE; (b) R2
    • Table 1. Mass fraction of Pb in fresh kelp samples

      View table

      Table 1. Mass fraction of Pb in fresh kelp samples

      SampleMass fractionSampleMass fraction
      10.273.0
      20.584.4
      31.095.0
      41.6106.0
      52.0118.0
      62.41215.0
    • Table 2. Comparison of multiple regression model performance indicators for different fusion methods

      View table

      Table 2. Comparison of multiple regression model performance indicators for different fusion methods

      Feature treatment methodModelTraining setPrediction set
      R2RMSE /(mg·kg-1ARE /%R2RMSE /(mg·kg-1ARE /%
      InterpolateRR0.9200.50616.930.9140.3619.32
      RFR0.9290.48616.770.9200.3409.83
      SVR0.9420.43913.280.9320.3229.15
      SelectRR0.9120.54716.640.9090.31011.08
      RFR0.9230.51416.530.9180.31310.03
      SVR0.9310.46913.790.9210.34610.65
      Feature equilibriumRR0.9630.35210.880.9340.3028.27
      RFR0.9650.32510.260.9440.2969.81
      SVR0.9680.3409.350.9570.2517.36
    Tools

    Get Citation

    Copy Citation Text

    Lingling Pi, Shengqun Shi, Weihua Huang, Zhiyong Ouyang, Junfei Nie, Jinling Xiao, Lianbo Guo. Spectral Feature Equilibrium for Enhanced Lead Detection in Kelp via LIBS-LIF (Invited)[J]. Laser & Optoelectronics Progress, 2025, 62(17): 1739023

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category: AI for Optics

    Received: Jan. 21, 2025

    Accepted: Mar. 17, 2025

    Published Online: Sep. 8, 2025

    The Author Email: Lianbo Guo (lbguo@hust.edu.cn)

    DOI:10.3788/LOP250557

    CSTR:32186.14.LOP250557

    Topics