Quantitative And Qualitative Analysis Based On Molecular Spectroscopy

Near infrared spectroscopy(NIR)and infrared spectroscopy(MIR)belong to the molecular spectroscopy,which can clarify the sample information of components and properties at the molecular level.At present,near infrared spectral analysis technology has been applied in the fields of petrochemical,pharmaceutical,food industry et al.It has achieved good benefits for economic and social.In addition,more and more portable Fourier transform infrared spectrometer(FTIR)correction method combined with multivariate quality control also has been applied to various fields,such as the quality of the paint industry and oil control.Spectral analysis can be used for quantitative analysis and qualitative analysis,however,whether it is near infrared and infrared spectroscopy,in the process of large-scale promotion,there are still some technical problems that have not been solved.Infrared spectroscopy is the most widely used and studied in all spectra.The infrared spectrum can reflect the vibration of the molecule and can be used to characterize and identify the species.In this paper,the related problems in qualitative analysis are studied:(1)In the quantitative analysis,to construct predictive correction model of multiple nature or the composition of unknown samples by chemometrics method and computer program,so as to achieve the quantitative analysis of nature or the composition of unknown samples.In practical application,the measuring instrument of change,the instrument to be measured and the influence of sample aging changes and other factors the establishment of a multi spectral instrument in Taiwan on the calibration model used in another instrument will have a large deviation,this paper can not even use.The spectral multivariate calibration model transfer research.This paper proposed an improved piecewise direct correction algorithm--SA-PDS method.The method for model transfer through PDS algorithm,using the spectral angle between(SA)as the criterion to select the transmission parameters,which makes the model transfer process is not affected by the properties of samples is not the reference value depending on the model,and can be two-way transfer.The tobacco near infrared data and infrared data for transfer of asphalt,determination of total sugar content in tobacco leaves and wax content of asphalt and model,and use the standard error of prediction(RMSEP)as the criterion for comparison.The results show that whether in the near infrared spectroscopy and infrared spectroscopy,transmission the prediction precision and prediction error model are smaller.(2)In the spectrum of qualitative identification,establish the relation between the spectra of different kinds of substances and their categories,according to the spectra of unknown samples can be predicted for its classification.In this paper,according to the main group of research.Qualitative analysis of different kinds of material are very similar because the material component is similar to that of the measured differences in their tiny molecular spectroscopy the use of spectral analysis technology combined with chemometrics,it is difficult to distinguish between differences in how to enhance the spectrum becomes the key to effective use of material spectral variables for classification and recognition.The image recognition method is applied to the qualitative analysis of the spectrum,the one-dimensional linear spectrum through the autocorrelation spectral image into two-dimensional.The visual difference spectrum enhancement,the image recognition method for classification and identification of the two-dimensional spectral images.In the classification and identification of 7 kinds of edible oil,Compared with PCA and KL-BP method,the image recognition method has higher prediction accuracy.