Removing Background of Raman Spectrum Based on Wavelet Transform

Raman spectrum is usually so weak that the noise usually distorts the interesting Raman bands. Especially, when the sample is focus, the strong fluorescence background often completely dominates Raman peaks. In this case, it is vital to remove fluorescence background for analyzing of Raman spectrum. Wavelet transform (WT) has proved to be a high-performance signal processing tool, and very efficient in removing low frequency noise. In this paper, WT was used to remove fluorescent background of Raman spectrum of Sudan I. Both the global thresholding method and the interval thresholding method were applied, and their de-noising performances were compared. The results indicated that the interval thresholding method is more efficient than the global thresholding method in removing fluorescence background of Raman spectrum.