Study On Milk Quality Detection Based On Fluorescence Spectroscopy

Milk,which contains abundant nutrition,plays a very important role in people's daily life.Milk is a primary source of nutrition for infant mammals before they are able to digest other types of foods,and appropriate consumption of milk is effective in promoting muscle growth.For the customers,the quality of milk is of great importance.Drinking unqualified or expired milk does a great harm to human body and even endanger human life,thus an appropriate method for milk quality detection is needed.The fluorescence of spectrum of particular material reflects the inherent information of it,and fluorescence spectroscopy technique is developed based on the principle.Fluorescence spectroscopy technology has a lot of applications in life,including food quality detection,classification,environmental pollutants detection,industrial production process testing,agricultural production,and medical beauty industry,etc.In this paper,the fluorescence spectroscopy technology,combined with chemometric data processing,modeling and analysis methods was used for milk quality detection.The F-2500 spectrophotometer was used to measure four kinds of milk,and the milk was classified by principal component analysis.In order to detect the freshness of milk,the fluorescence spectra of milk were obtained by using F-2500 spectrophotometer and a set of LED-induced fluorescence spectrum system.In the self-made system,the LED of 380nm peak wavelength excites fresh milk and the fluorescence spectra of different storage time are collected.The linear discriminant analysis(LDA)is performed on the milk spectral data and the regression model is established to predict the storage time of milk.The result shows that after 10 hours of unpacking the milk,the correlation coefficient between the regression prediction time and the true time of the milk storage is 0.988,which means the prediction is almost perfect.In order to verify the rationality of the model results,the bacterial experiment is carried out to test the bacterial content in milk at different storage time,which shows that the freshness of milk is consistent with the results of spectral detection.The result of this study shows that according to the different spectral of milk of different storage time,the real storage time of unpacked milk can be predicted accurately after 10 hours.