Liquid Electrode Glow Discharge Spectrum Analysis Algorithm And Optimization Desig

Chemometrics is an interdisciplinary subject which applies mathematics and computer science to chemistry,and spectral analysis is one of the important research contents of chemometrics.The core of spectral analysis technology is the spectral signal analysis algorithm,which directly affects the reliability of subsequent analysis and determines the analytical performance and application development prospects of spectral technology.Liquid electrode glow discharge spectroscopy analysis technology,as an emerging spectral analysis technology,has wide applications in fields such as rapid water quality analysis and detection.The qualitative and quantitative analysis of elements in solution samples can be realized by detecting the spectral peak and intensity information of elements.At present,the processing of liquid electrode glow discharge spectral information is mainly done manually,which cannot achieve rapid and accurate qualitative and quantitative analysis.In addition,existing spectral analysis algorithms have low peak resolution,insufficient accuracy,and poor adaptive ability,making it difficult to accurately obtain peak information.Therefore,this paper has carried out relevant research from the aspects of spectral analysis algorithm optimization and software analysis system design.The main research contents and contributions are as follows.1.Aiming at the problem that the existing spectral analysis algorithms have the weak resolution of the overlapping and weak peaks,a weighted continuous wavelet transform peak-finding method is proposed.In this method,Gaus2 wavelet with small linewidth is selected as the mother function of the transform.The scale parameter with small spacing is extremely important for the resolution of overlapping peaks and weak peaks.The user-defined threshold can effectively truncate the composite ridge and avoid the deviation of peak detection.The weighted continuous transform can compress and move the maximum value of the transform coefficient to the small-scale parameter range,and the maximum value in the small-scale range can more accurately reflect the peak position.Compared with other methods,the proposed method has obvious advantages in resolving overlapping peaks and weak peaks as well as peak finding accuracy,and has good peak finding reliability.2.Aiming at the problem of the poor adaptive ability of peak detection algorithms,a peak-finding method based on image processing technology is proposed.This method combines the image threshold segmentation method with the wavelet transform method.Firstly,the original spectral signal is transformed into the coefficients matrix by wavelet transform,and then the coefficients matrix is transformed into the gray image by the logical mapping mode.The gray image histogram is used to determine the image segmentation threshold,the gray image is segmented to separate the peak area,then the maximum value of the transformation coefficient is searched from the peak area,and finally the peak position is determined according to the ridge formed by the maximum value.Through the comparative analysis,it is found that the wavelet transform image segmentation peak detection method based on weighted continuous wavelet transform and Otus segmentation performs the best with better spectral peak recognition performance and lower false negative recognition rate.The proposed algorithm determines the threshold according to the image processing technology,which has strong adaptive ability and can be applied to the analysis and processing of various spectral signals.3.A software analysis system for glow discharge spectrum of liquid electrode with graphical user interface was designed to meet the practical analysis requirements of LEGD.The system consists of three modules: spectral signal acquisition,spectral analysis and quantitative analysis.It can realize the control of the spectrometer,the parameter setting of the spectrometer and the real-time collection and preservation of the spectral signal.It has good spectral pretreatment and spectral analysis effect,can accurately identify the qualitative display element information of the spectral line,and has the calculation ability of discharge parameters.Moreover,it can establish the calibration curve according to the concentration of the element and the corresponding spectral line intensity,and analyze the detection limit of the element and calculate the final quantitative results.The proposed analysis system can meet the analysis requirements of LEGD spectral analysis technology.