Raman spectral data denoising based on wavelet analysis

Abstract As one kind of molecule scattering spectroscopy, Raman spectroscopy (RS) is characterized by the frequency excursion that can show the information of molecule. RS has a broad application in biological, chemical, environmental and industrial fields. But signals in Raman spectral analysis often have noise, which greatly influences the achievement of accurate analytical results. The de-noising of RS signals is an important part of spectral analysis. Wavelet transform has been established with the Fourier transform as a data-processing method in analytical fields. The main fields of application are related to de-noising, compression, variable reduction, and signal suppression. In de-noising of Raman Spectroscopy, wavelet is chosen to construct de-noising function because of its excellent properties. In this paper, bior wavelet is adopted to remove the noise in the Raman spectra. It eliminates noise obviously and the result is satisfying. This method can provide some bases for practical de-noising in Raman spectra.