The Pretreatment Of Raman Scattering Signal And Its Application To The Statistical Pattern Recognition Of Medical Raman Screening Technology

The preprocessing of Raman signals is particularly important for the late analysis of Raman scattering signals.On the one hand,the Raman scattering signal is weak,and it is difficult to detect the scattered Raman signal clearly.On the other hand,Raman scattering signals are often interfered by some substances or external conditions,such as detecting machine noise,fluorescence interference,or Raman scattering signals of other substances in the sample.Therefore,it is necessary to preprocess some Raman scattering signals.First,the Raman light of pure R6 G is studied in depth.At the same time,some noise reduction algorithms which are widely used at present are introduced,and the simulation analysis of these algorithms is carried out.Such as wavelet transform,smoothing filter,empirical mode decomposition(EEMD)and minimum mean square error(LMS).The use of signal-to-noise ratio(SNR),root mean square error(RMSE)and correlation coefficient(?)to determine the level of noise reduction algorithm,the simulation results show that the LMS compared to the other three algorithms has a higher SNR.Secondly,through the study on SNR characteristics of Raman signal below 10 dB found that smoothing filter can well remove a lot of noise and low SNR than Raman signal,but there will still be a part of the noise can not be removed and the influence of the Raman signal details.The LMS algorithm has the characteristics of high efficiency and easy to implement,but the high algorithm has network instability,and the steady-state error and the convergence rate have unavoidable contradictions.Therefore,a new noise reduction algorithm for smoothing filtering combined with VS-LMS(variable step size LMS)is proposed.The experimental results show that compared to several traditional algorithms,the noise reduction effect has been greatly improved,and the spectral details are partially restored.Finally,through smoothing filtering combined with VS-LMS algorithm combined with multivariate statistical methods,we first discriminate thyroid dysfunction diseases which are now popular and neglected.The experimental results showed that the measured Raman signals using smoothing filter algorithm combined with VS-LMS pretreatment,and then through the PCA(principal component analysis)and LDA(multivariate statistical method)to enhance the signal of discrimination,discrimination the total correct rate reached 87.4%,indicating the feasibility using the multivariate statistical method of distinguishing method of thyroid diseases.