Study On Fast Inspect And Analysis Of Main Components Of Milk And Yogurt Based On Spectroscopy Technolgy

Dairy industry has made a significant development in recent years.Milk powder,which is essential to the infants,is one of the most common dairy products.Yogurt is worldwide regarded as one of the food that benefits the longevity of people,not only for its delicious taste,but also for its ability to resist senescence and cancer.With the dairy industry develops rapidly,some fraudulent phenomenon has largely increased.Certain kinds of inferior dairy products severely affect the growth of the infants,or even threaten their lives.Thus,we need an advanced determine technology for dairy industry.Since the traditional quality determination method which mostly bases on chemical technology can not meet the need of the society,it is indispensable to propose a fast,effective and nondestructive method to determine the quality of dairy products accurately.In order to determine the different components in dairy products,a series of criterion have been set up such as Rose-Gottlieb method for fat determination in mild powder,Kjeldahl nitrogen method and micro-direct distillation method for protein determination.However,these methods are too complicated to be popularized.Besides,determination methods for different components are of great diversities.Hence,to simplify the process of determination,a feasible approach is needed to detect different components of dairy products fast and precisely which can be realized by modern near infrared spectroscopy technique.According to the objective above-mentioned,the near infrared spectroscopy technology was used to discriminate the varieties and examine the interior qualities(fat,protein,calcium,iron,zinc,etc.)of the mild powder and yogurt.The main research contents and results are as follows:1.The short-wave near infrared spectroscopy technology was used to determine the brands and the fat,protein content of milk powder.Spectra of 350 milk powder samples in the region of800-1025 nm were measured.They were used to predict the brands and the components of milk powder.The brands were discriminated combined with LS-SVM and Fast ICA.The results showed that the ICA-LS-SVM model for milk powder brand discrimination reached an accuracy of 98%.In the meantime,the fat,protein contents were determined by LS-SVM and ICA-LS-SVM methods.These two methods both have satisfying results in predicting the main component contents of milk powder based on short-wave near infrared spectroscopy technology.The study demonstrated that the short-wave near infrared spectroscopy analysis can discriminate the brands and determine the main component contents of milk powder rapidly and reliably.2.The near infrared and middle infrared spectroscopy technologies were employed to the research on the determination of protein and fat contents of milk powder.LS-SVM was used for calibration.The study indicated that a better result can be achieved using MIRS region for protein content prediction,while the whole region is the best for fat content prediction.Both of the prediction accuracies were higher than 93%.These results proved that infrared spectroscopy technology can realize the fast and nondestructive determination of protein and fat contents of milk powder.3.The calcium,iron and zinc contents of milk powder were measured by near infrared and middle infrared spectroscopy technologies.The LS-SVM model were calibrated separately based on near infrared,middle infrared and whole region.The results indicated the feasibility of infrared spectroscopy technology for the fast and nondestructive determination of calcium,iron and zinc contents of milk powder.4.A new fast approach to discriminate varieties of yogurt based on near infrared spectroscopy technology was proposed.In this research,the spectra of five typical varieties of yogurt were measured by a spectroradiometer.Then clustering analysis was made using principle component analysis.Artificial neural network with principle components as input was used to establish discrimination model.The principle component analysis indicated that principle component 1 and 2has an accumulative reliability of 98.956% and accumulative reliability of the first 7 principle component is 99.97%.The accurate rate of variety fitting and prediction were both 100%.5.The pH of yogurt were determined fast and nondestructively by near infrared diffuse reflection spectroscopy technology.There were five typical brands and each brand had 27 samples.Firstly,the diffuse reflection spectra were taken.Then a pH meter was used to detect the pH values.The correlation between the spectral absorbance data and the pH of yogurt was built based on partial least squares(PLS)regression models.The correlation coefficient,standard error of calibration and standard error of prediction of the PLS-based model were 0.890,0.037 and 0.043.The correlation coefficient of prediction of the unknown samples was 0.918,which means the pH values of yogurt can be determined fast and nondestructively.6.An approach to determine the SSC(soluble solids content)of yogurt using near infrared diffuse reflection spectroscopy technology was proposed.There were five typical brands and each brand had 25 samples.Firstly,the diffuse reflection spectra were taken.Then Abbebenchtop refractometer was used to detect the SSC of yogurt.The correlation between the spectral absorbance data and the SSC of yogurt was built based on partial least squares(PLS)regression models.The correlation coefficient,standard error of calibration and standard error of prediction of the PLS-based model were 0.894,0.356 and 0.389.The correlation coefficient of prediction of the 35 unknown samples from five brands was 0.934,which means the SSC of yogurt can be determined fast and nondestructively by using near infrared spectroscopy technology.