Study On The Method Of Rapid Detection Of Meat Quality

At present,traditional methods are used to detect the quality of pork and other meats,which is destructive and can not meet the requirements of rapid,accurate,real-time and non-destructive in modern detaction.In this study,the fresh pork and beef were used as the research object.Near-infrared spectroscopy,polarization,paper chip technology and chemometrics applied used for the development of methods for the quick determination of the quality of meat.（1）Different parts of fresh pork（loin,front leg,hind leg）was applied for the quantitative analysis of muscle.The national standard methods were used for the detection of the shear force,cooked meat rate and pH of pork muscle samples,partial least squares is used to model（PLS）The established prediction model of pork tenderness showed that the number of latent variables（LVs）is 11,and the root mean squared error of calibration set（RMSEC）,root mean squared error of cross validation（RMSECV）and root mean squared error of prediction set（RMSEP）are 3.04N,5.87N and 6.99N,respectively,and the corresponding correlation coefficient of calibration set（R_C）,correlation coefficient of cross validation(R_CV)and correlation coefficient of prediction set（R_P）are 0.93,0.73 and 0.72,respectivelyhe established prediction model for cooked meat percentage showed that the number of LVs is 10,and the RMSEC,RMSECV and RMSEP are 1.13%,1.96%and 2.52%,respectively,and the R_C,R_CVV and R_P are 0.94,0.83 and 0.75,respectively.The established prediction model for pH showed the number of LVs is 6,and the RMSEC,RMSECV and RMSEP are 0.24,0.29 and 0.29,respectively,and the R_C,R_CVV and R_P are 0.80,0.72 and0.70,respectively.The results of wavelength selection showed that the RF-PLS model is better than UVE-PLS.The established RF-PLS model of pork tenderness showed that the number of LVs is 10,and the RMSEC,RMSECV and RMSEP are 2.78N,4.40N and 6.05N,respectively,and the R_C,R_CVV and R_P are 0.94,0.85 and 0.83,respectively.The established RF-PLS model for cooked meat percentage showed that the number of LVs is 10,and the RMSEC,RMSECV and RMSEP are 1.02%,1.48%and 2.59%,respectively,and the R_C,R_CVV and R_P are 0.95,0.90 and 0.77,respectively.The established RF-PLS model for pH showed that the number of LVs is 9,and the RMSEC,RMSECV and RMSEP are 0.20,0.26and 0.27,respectively,and the R_C,R_CVV and R_P are 0.86,0.77 and 0.75,respectively.（2）Different parts of fresh beef（loin,front leg,hind leg）was applied for the quantitative analysis of muscle.The established prediction model of beef tenderness showed that the number of LVs is 4,and the RMSEC,RMSECV and RMSEP are 14.29N,17.54N and 16.07N,respectively,and the R_C,R_CVV and R_P are 0.76,0.62 and 0.57,respectively.The established prediction model for pH showed that the number of LVs is 7,and the RMSEC,RMSECV and RMSEP are 0.25,0.32 and 0.25,respectively,and the R_C,R_CVV and R_P are0.82,0.69 and 0.62,respectively.The results of wavelength selection showed that the RF-PLS model are better than UVE-PLS.The established RF-PLS model of beef tenderness showed that the number of LVs is 4,and the RMSEC,RMSECV and RMSEP are 13.66N,17.17N and 16.03N,respectively,and the R_C,R_CVV and R_P are 0.78,0.65 and 0.59,respectively.The established RF-PLS model for pH showed that the number of LVs is 7,and the RMSEC,RMSECV and RMSEP are 0.22,0.27 and 0.27,respectively,and the R_C,R_CVV and R_P are 0.86,0.78 and 0.65,respectively.（3）The polarization-based near-infrared spectroscopy acquisition system was used for the first time to detect the pH and tenderness of beef.The effect of muscle texture on the model was discussed.The results showed that when the light direction is perpendicular to the muscle texture,the model results were the best.The influences of beef tenderness and pH model under four states with polarization angles of 0°,30°,60°and 90°were discussed.The results showed that the beef tenderness model with polarization angle of 90°is the best model,the RMSEC,RMSECV and RMSEP are 9.97N,16.46N and 17.31N,respectively,and the R_C,R_CV,and R_P are 0.90,0.71 and 0.69,respectively.The results showed that the beef pH model with polarization angle of 60°is the best,the RMSEC,RMSECV,and RMSEP are 0.20,0.40 and 0.46,respectively;the R_C,R_CV,and R_P are 0.91,0.66 and 0.63,respectively.The results of this work provide new methods and ideas for the quality test of the meat market in the future.（4）In this work,a method for detecting beta-galactosidase was established using paper chips.Whatman grade 1 cellulose chromatography paper was appleed as detection unit instead of traditional reagents.The optimum reaction pH is 7.0 and the optimal reaction time is 5 min.Paper chips were used to measure the RGB values with different activity ofβ-galactosidase at the optimum pH and the optimal reaction time.The regression equations in the three channels are obtained,and the R² are above 0.94,and the R² of the G-value fit curve reaches 0.98.Theβ-galactosidase activity was calculated to be 0.10 KU/mL based on the fitting curve of the total chroma of the G channel.The fitting results are good,which provid a new method for the rapid detection of E.coli on-site.