[An integrated on-line processing method for spectrometric data based on wavelet transform and Gaussian fitting].

Miniature mobile field spectrometry is pivotal equipment for qualitative and quantitative in-situ analysis of chemical substances. To solve the problem of spectrum signal interfered by complicated noise, overlapped and irregular peak shape recognition, and quick monitoring, an integrated on-line processing method for spectrometric data based on wavelet transform and Gaussian fitting was developed. In this way, toluene and perfluorotributylamine were processed, and the results shows that the integrated method can powerfully and effectively eliminate the noise, retain the original feature, and correct the overlapped and asymmetrical peaks, which can improve the analysis accuracy of instrument, and also achieve data compression. In addition, the method satisfies the requirement of on-site analysis for mobile field spectrometry. For the processing of mass spectra of toluene, at the characteristic peaks of 91 and 92, the SNR increased 1.3 times compared to that of moving average smoothing method, while the error between original peaks and theoretic peaks decreased 3.6 times. In addition, Gaussian fitting described the multipoint mass spectra data by three Gaussian parameters, and achieved data compression. For the processing of mass spectrogram of perfluorotributylamine, the ratio of compression was 197 : 1.