Research On Key Technologies Of Raman Spectroscopy Pretreatment

In terms of substance detection,Raman spectroscopy is widely used in various fields for its advantages of simplicity,rapidity,repeatability and nondestructiveness.With the further development of laser technology and nanotechnology,Raman spectroscopy technology is becoming more and more mature.However,in the acquisition of Raman spectra,noise interference may occur in every process from the emission of the spectrometer to the receiving end,and the presence of fluorescence will make the measured spectra have a great background,affecting the further analysis of Raman spectra.Therefore,it is of great significance to preprocess the acquired spectral data.This thesis first introduced the background,significance and research status of Raman spectroscopy,as well as the principle of Raman effect,the structure of Raman spectroscopy and the application of Raman spectroscopy.Then the key and difficulty design of the design are analyzed according to the structure of spectrometer.Next,the possible sources of noise in each process obtained by Raman spectral data are deeply explored.The noise types at the receiving end are mainly introduced.According to the source of the noise,the windowed median filtering method,the SG filtering method and the wavelet threshold filtering method are used to denoise the measured Raman spectrum under various parameters.After comparing the best effects,the superiority of wavelet threshold filtering is confirmed.Then an improved method,piecewise polynomial fitting,is proposed on the basis of polynomial fitting for baseline correction.The spectral data is segmented and then the proper order is fitted respectively.The iterative optimization method is used to eliminate discontinuities between piecewise points.In the verification of the algorithm,the simulation spectrum is constructed to introduce the whole process of the algorithm implementation.Compared with traditional polynomial fitting and piecewise linear fitting in real spectrum,the proposed baseline correction algorithm is proved to be effectiveFinally,the Raman spectra of various samples were measured,and the preprocessing process including noise suppression,baseline correction,normalization and peak identification was carried out.The effectiveness and limitations of the improved method were further analyzed.