Automatic characteristic spectral line selection method used for laser induced breakdown spectroscopy analysis

Abstract

The invention belongs to the field of spectral analysis. Automatic and fast selection of multiple characteristic spectral lines related to specific element content in massive spectral data can be realized without artificial determination processes; selection efficiency can be greatly improved; the number of characteristic spectral lines participating in modeling can be effectively reduced; calculating processes can be simplified; the steadiness and accuracy of a quantitative prediction model can be enhanced; and a verification method is also provided. An automatic characteristic spectral lineselection method used for laser induced breakdown spectroscopy analysis includes the following steps: step one, performing first full spectrum correction on spectral raw data X; step two, performing second full spectrum correction on the spectral raw data X' corrected by the step one; step three, rejecting unimportant spectral line variables; step four, a step of iteration, namely the step of performing partial least squares analysis on the residual variables selected by the step three; and step five, a step of determination, namely the step of determining that cyclic residual spectral lines are the final selection results. The method is mainly applied to spectral analysis occasions.

Description

technical field [0001] The invention belongs to the field of spectral analysis, in particular to a method for automatically selecting characteristic spectral lines for laser-induced breakdown spectral analysis. Background technique [0002] Laser-Induced Breakdown Spectroscopy (LIBS), or Laser-Induced Plasma Spectroscopy (LIPS), is a new and promising analytical technique in atomic generation spectroscopy. LIBS technology uses high-energy-density pulsed laser as the excitation light source. The laser is focused on the sample surface to form plasma in the local ablation area, and emits radiation to the outside world. After a certain period of continuous spectrum, the atomic emission related to the sample composition is generated Spectra, which can now be collected by a fiber optic collection device and analyzed by a spectrometer. Compared with the traditional atomic emission spectrometry technology, LIBS technology does not require complex processing of samples, and can real...

AI Technical Summary

Problems solved by technology

Commonly used methods for automatic selection of characteristic lines include genetic algorithm (GA) and continuous projection algorithm (SPA), but these two algorithms are often time-consuming and inefficient, and the selection time can even reach several hours, which cannot be achieved in real-time detection. get applied

Images