Identification Study Of Milk Quality And Metamorphic Grade Based Spectroscopic Techniques

In the characterization of milk, the main method is to detect how much the contentof the main components in milk, or to detect major nutrients share in the total percentageof fresh milk, protein and fat proportion is the highest, so often detected as two importantindicators. Currently, in the detection of milk ingredients, the traditional chemicalanalysis methods are still have the dominant position, but this method is slow, costly andeasy to destroy the sample and other shortcomings, it is difficult to implement on-line,real-time monitoring on products. To overcome these shortcomings of traditionalmethods, using near-infrared spectroscopy is a good choice, near-infrared spectroscopyhas unparalleled advantages in milk quality identification, the method detect fast, do notneed to do too much pretreatment on samples, can achieve non-destructive measurements.This paper attempts to use near infrared transmission spectroscopy to detect the proteinand fat content of milk, providing two important basis for milk quality identification. Themain research work includes the following aspects:1. Establish fresh milk protein and fat content quality testing model in differentspectral forms and in different spectral bands. This paper study on90samples providedby the Chongqing Tianyou milk group, using Omni-λ3007spectral collection systemprovided by Instrument Co., Ltd. Beijing Zhuoli, to detect transmission spectrum of thesample, then transform the original spectrum into three new forms of spectral data withmathematical transformation, that is the energy spectra, proportion spectra andabsorbance spectra. Modeling the prediction model of protein and fat proceed from thethree spectra of the samples separately, selected optimum prediction model by comparingthe predictive ability of calibration model of each spectral data, based on this, discussedthe model predictive ability influenced by the wavelength selection and he modeltransform form of spectra. The results showed that the predictive model of protein and fatcreated by the angle of absorbance is best, the correlation coefficients (R2)were0.9984and0.9985, the squared errors were4.775106and6.536106, RMSE were5.463104and6.392104.2. Established a determination model of milk degree deterioration based onspectroscopy. Based on the study on milk quality accreditation, applied thecharacterization model to the practical application of the milk deterioration extent determine in-depth. In this paper, a prediction model has be established based onnear-infrared transmission spectrum. The results show that the quality of milkidentification system based spectroscopy can be well for determining the degree of milkdeterioration, it is good at distinguishing the four important stages of milk qualitativechange, that provides a new idea for food safety testing technology.The results show that: the model established by absorbance method combines PLSis better than the method of energy and proportion; the model of protein established by2180nm is better than2056and the model of fat established by2190nm is better than2150nm. The correlation coefficient of the protein model and fat model were increasedby0.0234and0.0034. The method used in this paper to detect the main ingredients offresh milk can get higher accuracy, It means this method have a higher accuracy thanconventional near-infrared quantitative analysis, meet the rapid detection of milk qualityaccreditation requirements on the market, The detection system can also use to determinethe degree of milk deterioration, and has a low cost, easy operation.