| Near-Infrared Reflectance Spectroscopy(NIRS)technology is an increasingly mature detection technology since the 21st century.Its application is simple and efficient,and it has a very broad application prospect in the field of meat detection.The purpose of this study was to use near-infrared spectroscopy to rapidly and accurately detect the values of various physical and chemical indicators in lamb,identify lamb varieties and origins,predicting its shelf life and comprehensively assessing the quality and safety of lamb.This paper includes the following three experiments:(1)The purpose of this experiment was to use WINISI III software and Foss Calibrator software to establish predictive models of 6 chemical components(MT,EE,CP,GLU,Ash,P)of 203 lambs respectively.The WINISI III software calibration results show that the RSQv of the MT,EE,CP,GLU,Ash,and P predictive models are:0.86,0.93,0.75,0.61,0.57,0.41 respectively;the Foss Calibrator software calibration results show that the R~2 of the MT,EE,CP,GLU,Ash,and P predictive models are:0.84,0.87,0.83,0.53,0.51,0.33,respectively.The conclusions obtained by the two calibration software are basically the same,indicating that the predictive models of MT,EE and CP can be applied in actual production;the predictive models of GLU and Ash can be used for rough analysis and screening,but need to be further optimized;the predictive model of P is less relevant and cannot be applied in actual production.(2)In this experiment,148 Hu sheep,62 Australian-Hu F1 lambs,and 84 Dorper sheep were collected.A total of 294 samples were used to construct OPLS-DA and O2PLS-DA identification models of different lamb breeds.395 Hu sheep samples were collected to construct OPLS-DA and O2PLS-DA identification models of lamb from different origins(112 samples from Wuwei Pukang,148 samples from Qingyang Zhongsheng,and 135 samples from Shandong Runlin).The results of the variety identification model showed that the total discriminant accuracy rate of the calibration set of the OPLS-DA and O2PLS-DA models were 98.30%and 99.15%,respectively;and that of the validation sets were 98.31%and 100%,respectively.The results of the origin traceability model showed that the total discriminant accuracy rate of the calibration set of the OPLS-DA and O2PLS-DA models were 97.15%and 100%,respectively;and that of the validation sets were 96.20%and 98.73%,respectively.Whether it is a discriminant model for different breeds of lamb or a discriminant model for the same breed of lamb from different origins,the total discriminant correct rate of the O2PLS-DA model is higher than that of the OPLS-DA model,the O2PLS-DA models are their best discriminant models,respectively.(3)In this experiment,110 neck samples of Small Tail Han sheep were collected,and the purpose was to use NIRS technology to establish a quantitative model of physicochemical parameters(MT,CP,EE,L*,a*,b*,p H,TVB-N)of lamb samples stored for 0-10 days;and try to establish zero-order and first-order kinetic models for each index;The spectral-dynamic model and the near-infrared spectral discrimination model of refrigerated lamb were constructed respectively to predict the storage period of refrigerated lamb;and then combined with spectral clustering results and sensory evaluation to comprehensively evaluate the quality and safety of lamb.The results showed that the RSQv of each predictive model for lamb MT,CP,EE,L*,a*,b*,p H,and TVB-N were:0.94,0.87,0.61,0.63,0.68,0.75,0.72,0.86,respectively.Through comparison and analysis,only the TVB-N index conforms to the change rule of the food kinetic model,and its zero-order kinetic equation is A(t)=8.54+1.895t.The shelf life,freshness and threshold value of refrigerated lamb can be accurately predicted through the spectrum-dynamic model and near-infrared spectral discrimination model of TVB-N index.Using spectral clustering results combined with sensory evaluation,the lamb refrigerated for 0-10 days was roughly divided into five categories:fresh 0-1d;second-fresh 2-5 d;early stage of deterioration 6 d;middle and late stage of deterioration 7-9 d;spoilage 10 d.To sum up,this study used NIRS technology combined with chemometric methods to construct a quantitative prediction model of various physical and chemical indicators of refrigerated lamb,a qualitative identification model of variety and origin,and a spectro-kinetic model.It provides a theoretical basis for quickly evaluating the physical and chemical quality of lamb,identifying varieties and origins,assessing quality and safety,and predicting its shelf life. |
