Research Of Detecting Technology For Nutrition In Milk By NIR

With the development of the standard of living, no more have people beensatisfied of the abundance and amplitude of food, and are much moreconcerned about nutrition of food. As a new subject, chemometrics isdeveloped greatly. As an important means in chemometrics, NIR is advancedenormously accordingly. In this paper, the milk which is one of the peoples'main drink is measured online. The main research works are as follows:(1) Based on the analysis and synthesis of present milk measurement andits developing tendency, the paper proposed that applying near infraredspectrum technique and optical fiber to on-line measurement technique of thenutrition in milk.(2) Based on the fundamental theory of NIR spectrum analysis, thespectral signature of the nutrition in milk is researched.(3) Starting from the digital fourier filtering pretreatment and wavelengthselection by genetic algorithms in regression models were discussed.(4) Secondly, The research analyzed the configuration of the fiber-opticNIR measuring system, studied the spectrum property of light source, thetransmission property of fiber-optic, the spectrum response property ofphotoelectric detector and the precise coupling method of light source withoptical fiber, built the mathematics model of the all fiber detection system,selected the optimal wavelength for milk detection.(5) At last, The research of noise resource and noise control method forthe NIR detection system and improvement of the system stability andreliability.The experiments shown that the system has many merits, such as highstability, insusceptible by environmental factors, high speed and high degree ofaccuracy, it can measure multi-component.