Laser Induced Breakdown Spectroscopy Data Processing Method Based on Wavelet Analysis

摘要

In this paper, we present a data processing approach for Laser induced breakdown spectroscopy (LIBS). This method is based on wavelet analysis and pattern matching. First, it uses wavelet transforms to decompose the laser induced spectrum data which comes from the sample and obtain the decomposition coefficient of spectrum, then reconstructs the feature background spectrum by means of low frequency coefficient. Through using pattern cluster method to divide the spectrum data of calibration sample into some subsets, then do the calibration for each spectra data in each subsets. Second, we extract effective measurement pattern class template and calibration parameter from the spectrum subset which has the minimum differ between the result of calibration sample and the reality value. In practical process of measurement, we use effective measurement pattern class template to match the spectra data to identify the effectiveness of the measurement. Therefore, we can calculate element contents with the calibration parameter achieved before. This method can decrease the times of laser excitation and increase the measurement accuracy effectively.