Signal Processing And Software Implementation For Dynamic Raman Imaging System Of High Spatio-temporal Resolution

Dynamic Raman imaging system of high spatio-temporal resolution is the Ra-man microspectroscopy imaging system designed for monitoring of electrochemical reactions in situ.In order to improve the sensitivity of the system,sigal processing methods must be adopted.In addition,in Raman imaging aspect,in order to improve the temporal resolution,we often use shorter scanning time in the experiments.In this condition,for getting elaborate Raman images in real-time,fast Raman imaging algorithms must be developed which could reduce the computational cost and sig-nificantly improve the quality of Raman images.This dissertation aims at exploring relevant signal processing algorithms which possess good performance in the low SNR condition of the dynamic Raman imaging system and programing relevant software.The main work of this dissertation includes:(1)Proposed a Raman signal reconstruction algorithm based on matching pur-suit for the single point spectrum of the Raman scanning imaging spectra datasets,which achieved good perfomance in the baseline correction and de-noising com-pared to the traditional algorithms.It first finds the local maxima and minima points in the mean spectrum,and divides the nosie spectrum into with-signal bands and no signal bands according to these extreme points.In the next,matching pursuit is used to approximate the true Raman signal in the with-signal bands and finally recon-struct the whole signal.(2)Proposed 2 signal processing algortihms for fast Raman imaging:namely,fast Raman imaging algorithm based on the sparse approximation of characteristic peak(CPSA),fast Raman imaging algorithm based on FFT and IFFT.The two algo-rithms are both univariate fast Raman imaging methods,which can achieve better imaging results than conventional algorithm,and with less computational complexi-ty and faster imaging speed.(3)Utilized the combination of C#and MATLAB languages to program good mannerd software to realize these functions such as baseline correction,spectral de-noising and Raman imaging.This software could embedded in the dynamic Ra-man imging system,also can be utilized independently or embedded in other Raman imaging systems.Experimental sigals were used to verify the effectiveness and accuration of the proposed signal reconstruct algorithm,and comparsions with conventional algorithm were presented together for reference.Experimental Raman spectra data sets col-lected from Au-coated silicon with clear physical structure were used to verify effec-tiveness and accuration of the proposed fast Raman imaging algorithms.Addionally,an evaluation method of Raman image quality also have been proposed creatively.