Based on the multi-spectral imaging fresh pork shelf life prediction research

pork is rich in nutrients, is essential to human life and food source. In recent years, China's pork production continue to improve, people increase the demand for pork products, also put forward higher requirements for its quality. However, because the fat in meat, protein content is rich, water activity is high, in the process of processing, storage, sale, it is easy to microbial contamination and influenced by environmental factors, the product deterioration, which lost its edible value. Fresh pork by endogenous enzymes in circulation, storage process, the external environment, the effect of microorganism and decaying. The protein in fresh pork under the action of enzymes and bacteria, decompose and produce the alkaline nitrogen content, such as ammonia and amine and combined with acidic substances within the organization, formation of salt, the ground state n. TVB in fresh pork N content is with putting the extension of time, showing increased slowly, steadily increased and increased dramatically change the three process.

spectral imaging technology is a new platform technologies. It is spectral analysis technology and the combination of the image analysis technique, spatial resolution and spectral resolution, can access to the object's spatial and spectral information at the same time. Spectral imaging technology in the application of food, agricultural products is the widely, Peng and Lu research using spectroscopic scattering properties prediction beef pH, tenderness, and color. Multi-spectral imaging technology in portable, small batch production has more advantages.

using nonlinear regression method, with lorentz function containing the four parameters scattering curve fitting of various wavelengths. Contains four parameters of lorentz shown by the following functions: including: R for scattering curve on any point of the reflected light intensity ( Grey value) : z for scattering distance beam incident point at a distance, Circle radius) ； A is scattering curve of progressive values; B for scattering curve peak: c for half wave scattering curve bandwidth; D for scattering curve slope of turning points.

to the scattering of individual wavelengths do LD function fitting curve, so scattering images of individual wavelengths can use four parameters of LD function to describe, further analyze the parameters may predict TVB - fresh pork N。 As shown in the image of a figure such as scattering curve fitting. To build up the forecast model on the 4 volatile base nitrogen partial least-squares regression ( PLSR） Method because of its principal components are given and the correlation of components to be analysed, built by linear model usually has higher prediction precision, so this article USES the PLSR method to establish the fresh pork TVB The prediction model of N. Samples can be divided into two groups, the sample number for multiple of 3 7 samples for prediction set, the remaining 14 samples for calibrating. Each sample of TVB N reference value corresponding to the four parameters of the LD function of the scattering. Calibrating the 14 samples of TVB N reference value and 392 ( 14 x7x4) A partial least squares regression parameter values do. Model prediction correlation coefficient reaches 0. 87, standard error is 2. 50 mg / 100 g, for the PLSR method modeling prediction effect.

2。 5 shelf life prediction model is established to observe the change of the reference value, can be found in this study, TVB N a logarithmic growth as time changes, changes in rapid increase stage, so this article USES the logarithmic function fitting of TVB N curve, based on the shelf life prediction model of fresh pork.

rate parameter; T is time ( d） 。 To ( 2) Type transposition, the exponential deformation processing of type: type ( 3) In type ( 2) Can be calculated, derived on the basis of the shelf life of fresh pork.

TVB - N the change rule of a logarithmic characteristics, fitting correlation coefficient of 0. 93, the standard error of 1. 76mg/100g. Fitting effect as shown below. Three parameters fitting to A = 29. 1076，B=19. 8618年,k = 0。 1824. Will the three parameter substitution ( 3) Type, set up the shelf life of fresh pork ( 货架寿命,SL, d) The forecast model of: TVB N change law with time of volatile base nitrogen prediction model to predict 21 TVB The value of N in ( 4) Type, the right to obtain corresponding specimens of shelf life value, for the prediction of the shelf life of fresh pork, rendering can be seen that predict the overall trend is correct, but poor precision.

3。 Conclusion the shelf life of fresh pork is one of the important basis to measure the value of goods, therefore, to get a sense of fresh pork shelf life nondestructive, rapid and reliable detection method is very important. Traditional methods of chemical laboratory analysis, although has the advantages of high accuracy, reliable, but its analysis process cumbersome and time-consuming, the sample is destructive and gradually can not meet the requirement of the rapid development of production and economy. Spectral image technology is a combination of spectrum and image analysis technology, possesses the characteristics of fast, nondestructive and multi-spectral imaging technology in portable, small, batch production has more advantages. Multispectral imaging technology is presented in this paper, on the basis of in view of the small sample, nonlinear problems, to explore feasible detection methods for the shelf life of fresh pork. The results show that using the multi-spectral imaging technology combined with the corresponding mathematical modeling method is feasible for forecast the shelf life of fresh pork.