Study On The Parameters Of Mutton Freshness By Near Infrared Spectroscopy

Mutton is the main consumption of meat products in the northwest minority areas of our country,with the diversification of people's dietary structure and the continuous improvement of the nutritional value of mutton,mutton has shifted from some people's consumption to national consumption.Chilled mutton is more mature,more nutritious and taste better than frozen mutton and hot fresh mutton,which has become the mainstream of consumer spending.However,due to China's cold chain system is not perfect,chilled mutton in the processing,storage,transportation and sales are prone to deterioration,internal components and external senses will change.In different packing ways,the breathing patterns of meat are different,the internal components and external sensory characteristics change differently,and the influence on the shelf life and quality changes of meat is also different.PH,color brightness(L*)and the total number of microbial colonies were closely related to the freshness of meat,and the three were important indicators to evaluate the freshness of meat.Traditional detection methods have been unable to meet the requirements of modern large quantities,fast and non-destructive detection of meat,which seriously restrict the development of Xinjiang's chilled mutton go high-end devdlopment line.In this paper,the chilled mutton of Xinjiang is taken as the research object,the influence of fiber structure parameters and acquisition mode on the spectrum is analyzed by experiment and summarizing the domestic and foreign scholars' research results.Near infrared reflectance spectroscopy and chemometrics were used to detect the pH,L* and total bacterial count of vacuum packaged chilled mutton under different storage time.Near infrared reflectance spectroscopy and chemometrics were used to detect the pH,L* and total bacterial count of vacuum packaged chilled mutton under different storage time.Comparing the results of different spectral pretreatment methods and the models built in different bands,the best feature wavelength extraction method and the best quantitative prediction model for the fresh parameters of vacuum packed chilled mutton were optimized.To provide theoretical support and technical support for the rapid and nondestructive detection of fresh mutton quality.The main contents and results are as follows:(1)The working principle of the optical fiber sensor is studied,summarized and analyzed the domestic and foreign scholars on the research results of the influence of the optical fiber structure parameters on the light intensity modulation and combined with experiments.The principle of selection and design for the acquisition of near-infrared spectroscopy of mutton was determined,and the "9 in 3 out" coaxial A type fiber and the "3 out" coaxial B type fiber matched with the VIVO light source were completed under the guiding principle design.(2)Based on VIVO light source,QP400-1-VIS-NIR,QR400-7-VIS-NIR fiber and NIRQuest256-2.5near infrared spectrometer composed of spectral acquisition system,the influence of different acquisition parameters,instrument detection time and acquisition distance on the quality of spectral acquisition was analyzed.The longer the integration time,the greater the reflection value of the spectrum,the better the signal to noise ratio,but the integration time is too long lead to white reference spectrum saturation,causing spectral part of the band deformation.Increase the smoothness,the spectral curve is smoother,the noise is reduced,and the smaller crest trough is lost.Increase the average number of times,the quality of the spectral signal is increased,but the error of the system is increased.The longer the instrument's detection time,the more serious the spectral baseline drift.With the increase of the distance between the optical fiber probe and the sample,the spectral reflection intensity increases first and then decreases.(3)In 1 to 20 days of storage time,the pH value of vacuum packaged chilled mutton decreased first,then increased and then decreased.The effects of different pretreatment methods on the PLSR model are compared and analyzed,and the optimal spectral preprocessing method is the combination of 1 order derivative(1D),15 points S-G smoothing and centralization(MC).A comparative analysis of the differentbands of W-PLSR,GA-PLSR,SPA-MLR,GA-SPA-MLR,SiPLS-PLSR model.It is found that the MLR model based on 15 characteristic wavelengths extracted by GA-SPA method is the best,the correlation coefficient Rc,Rcv,Rp were respectively 0.92,0.89,and 0.91,and the root mean square error of RMSEC,RMSECV,RMSEP were respectively 0.13,0.15,and 0.13.(4)Vacuum packaging chilled mutton color brightness(L*)first increase and then decrease and then increase in 1 to 20 days of storage time.The corresponding spectral best pretreatment method is a combination of 1 order derivative(1D),17 points S-G smoothing and numerical centering(MC).GA,SAP,GA-SPA and SiPLS were used to extract the characteristic wavelengths to establish the corresponding PLSR model and MLR model and compared with the full band spectral PLSR model,the MLR model of the 18 characteristic variables extracted by GA-SPA was the best,the correlation coefficient Rc,Rcv,Rp were respectively 0.95,0.93,and 0.91,and the root mean square error of RMSEC,RMSECV,RMSEP were respectively 1.36,1.62,and 1.91.(5)The total number of colonies of vacuum packaged chilled mutton decreased first and then increased in storage time.2 order derivative(2D),13 points S-G smoothing and numerical centralization(MC)combination method was the best pretreatment method for spectrum.By comparing and analyzing the effects of W-PLSR,GA-PLSR,SPA-MLR,GA-SPA-MLR and SiPLS-PLSR in different wavelength bands,which concluded that GA-SPA-MLR model(26 points)prediction effect was the best,the correlation coefficients Rc and Rcv were respectively 0.95 and 0.92,the root mean square error of RMSEC and RMSECV were 0.45 and 0.59,respectively.(6)Compared with GA,SPA and SiPLS feature extraction methods,the GA-SPA method can effectively eliminate the redundant information in the spectrum and obtain the characteristic wavelengths related to the physical and chemical values,and can reduce the modeling wavelength points while maintaining the accuracy of the model has no significant difference or slightly increased.