| Laser-induced breakdown spectroscopy(LIBS)is a novel spectroscopy technique that uses high-energy laser pulses to act on a sample to generate plasma for the atomic emission spectroscopy.The method enables physical and spectroscopic analysis of the material composition and content of the sample.LIBS spectroscopy has the advantages of simultaneous analysis of multiple elements,non-destructive(or micro-damage)detection,and thus has been widely used in various fields.The original LIBS spectra usually contains a large number of elemental characteristic peaks,and the accurate identification of these peaks is the prerequisite and foundation for the elemental detection and analysis.In this paper,in view of the shortcomings of low accuracy and lack of reliability of traditional spectral peak element identification methods,a new method of automatic identification of element spectral peaks is proposed.The method not only achieves the recognition of the peak elements of the spectrum,but also performs the background correction of the spectrum and the decomposition of overlapping peaks.The main research contents are validated according to the relevant quantitative experiments.The main work of the dissertation is as follows:1.Due to the influence of laser energy,gate width,delay and experimental environment,a continuous background will inevitably occur in LIBS spectrum.Considering that there is a certain correlation between the continuous background and the minimum point in the spectrum,a method of detecting and correcting LIBS continuous background based on spline interpolation is proposed.The interpolation calculation was performed between the appropriate minimum points in the spectrum of the proposed method and can reasonably estimate the continuous background in the spectrum and correct it.The method is validated through the calculation of signal-to-background ratio and the result of quantitative experiment.The results show that: the traditional method obtains the signal back ratio in the range of(10.09,25.63),and the interpolation method can get 26.96.Through the quantitative analysis results after background correction,the correlation obtained by the traditional method is range from(0.9134,0.9790),and the interpolation method achieves 0.9985.2.An error compensation method is proposed to correct overlapping peaks due to the overlapping interference between the spectral lines caused by the characteristics of the element spectrum and the influence of the experimental equipment and the environment during the collection of the LIBS spectrum.The validity of this method is verified by the calibration of the overlapped peaks of Cu-Fe in the spectra and the quantitative experiment of Cu.The results show that: compared with the traditional method,the overlapping peaks correction method based on error compensation can effectively improve the correlation about 0.86%-1.82% and reduce the fitting error about 18.12%-32.64%.3.The characteristic peaks in LIBS spectrum were studied,and provided a method to identify the corresponding elements of these peaks.The experimental spectrum was fitted first of the method to obtain the characteristic peaks in the spectrum,and extracts the center wavelength,the light intensity,the full width at half maximum and the centroid of the peaks as the peak characteristic parameter vector.The similarity analysis of the spectral peak feature parameter vector between the spectrum to be identified and the spectrum in the standard spectral line database is performed to realize the automatic identification of the corresponding elements of the spectral peak.The simulated spectra created by the NIST standard database and the spectra obtained from tea samples were used to experimental research.The results show that: the traditional correlation analysis method achieves 79.79% accurate recognition rate.The spectral peak recognition method described in this paper can get the accurate recognition rate about 87.88%,which verifies the validity of the method for the peaks identification of LIBS. |
PDF Full Text Request