Raman Spectrum Signal Processing Based on Wavelet Denoising Methods

Raman signal of carbon tetrachloride was collected by laser Raman spectrometer. As there is a large amount of spectral data, interference information and effective information coexist. These are not conducive to the spectral data for qualitative and quantitative analysis. Therefore, wavelet thresholding method was used to the original Raman spectroscopy denoising was used. The results show that wavelet hard-threshold can obtain the optimal denoising quality. When wavelet base function being db2, scale decomposition being 4, the threshold quantization being ‘Heursure’, hard-threshold value being processed, signal to noise ratio(SNR) of the reconstructed spectral is the maximum, and root mean square error(RMSE) is the mini-mum. It showed that wavelet hard-threshold could effectively remove the noise information of the raw Raman signals of carbon tetrachloride, and saved the spectra details to the maximum extent.