The Application Of Wavelet Denoising Method In Spectrum Data Processing

During the spectral measurement, many aspects such as the instrument itself, externalenvironment and the users can affect the measurement data results. The aspect mentionedabove will affect the accuracy of the measured spectrum data inevitably contains noise.Through to deal with the noise, the influence of noise can reduce, the accuracy of theinstrument can be improved and also the reliability of its application can be raised.Wavelet analysis in signal analysis play a strong role, it can be used to processing thesignal in space domain and the time domain for effective localization analysis. It hasmulti-scale wavelet analysis (resolution) characteristics and can be used to observe the signalin easy and complicated way. At the same time, the wavelet analysis also has thecharacteristics of constant relative bandwidth. Choose a suitable wavelet method, the localdomain characters can be analyzed in time domain and frequency domain; it is very favorablefor the special signal analysis.This paper first introduced the basic concept of the wavelet analysis, including thecontinuous wavelet transform, the discrete wavelet transform, the multi-resolution analysistheory and MALLAT algorithm, etc. Then, with the wavelet threshold, the wavelet domainWIENER filter and Bayes wavelet based on the statistical model, the wavelet denoisesprocess method on simulation data for the spectrum is described. According to the scale of thesignal-to-noise ratio before and after processing, the wavelet threshold, that is better than theother two denoising method, is selected to de-noising the measured spectrum data captured bya real CCD camera in spectrum.This paper constructed a transfer function for CCD spectrometer; the study of the dataprocessing method with transfer function is described and this method was used for CCDspectrometer measured data processing on the base of simulation analysis. By this method ofprocessing ruby fluorescence spectrum, the absorption peaks and the stimulating peak can beobserved clearly. In order to adapt to different spectrum curve treatment, the piecewisefunction processing method is used for each different interval of the ruby fluorescencespectrum for the transfer function processing. The experimental results show that this methodcan further improve the accuracy of the instrument and resolution, and it can make it possiblefor the wide application of CCD spectrometer.