| The traditional methods of mutton quality detection and storage time prediction have disadvantages such as high price,long time-consuming and damaged samples.This study was based on the Yanchi Tan sheep of Ningxia,with two storage temperatures of cold fresh(4?)and room temperature(15?),the hyperspectral images(400?1000 nm,900?1700 nm)of mutton were collected at different storage time and temperature.At the same time,the relative color(L,a,b),pH value,moisture content,volatile base nitrogen(TVB-N)content and shear force value of the corresponding color and lustre(L,a,b)were measured by color difference meter,pH meter,water rapid measuring instrument,disease meat detector and texture analyzer.By comparing the traditional dynamic model and the correlation coefficient of hyperspectral model,the quality indexes of the time varying dynamics and the hyperspectral model were selected,and the dynamic model based on the spectral eigenvalue was established to predict the quality change and storage time during the storage of mutton.It can truly achieve fast,efficient,non-destructive and accurate.The main results are as follows:(1)During storage,the pH value of mutton first decreased and then increased,and the color(L,a,b)increased first and then decreased,and the final values of L and b were lower than the initial values.The final value of a was higher than the initial value.The moisture content increased first and then decreased and eventually tended to slow.The TVB-N value increased gradually.The shear force kept decreasing.Cold freshness was consistent with room temperature storage quality,but progressed more slowly.(2)The dynamic model of each index was established,and the better quality indicators were compared and analyzed.The results showed that:the correlation coefficient of the zero-level model of the pH value of mutton being stored at room temperature was 0.9120,kinetic equation was A(t)=5.188+0.069t;the correlation coefficient of TVB-N value of mutton at room temperature was 0.9880,kinetic equation was B(t)=0.489e0.169t;the correlation coefficient of zero value TVB-N model for chilled mutton was 0.9810,kinetic equation was C(t)=6.557+1.343t.(3)The Hyperspectral Information of mutton collected at different storage temperatures was used as the input value,and the hyperspectral model was established by combining the pH and TVB-N indexes of the dynamic model.The results showed that:the SG-S(3,7)+De-trending(4)-PLSR model of pH value of mutton stored at room temperature under 900?1700 nm was the best,and the prediction correlation coefficient was 0.8187.The SNV-PLSR model of pH value of mutton in cold fresh storage of mutton under 400?1000 nm was the best,and the prediction correlation coefficient was 0.9343.The original spectral PLSR models with TVB-N value of 400?1000 nm at room temperature and cold and fresh storage were the best,the prediction correlation coefficients were 0.8767 and 0.8436.(4)Based on the pH value of mutton at room temperature,a hyperspectral binding kinetic model was established,and got the prediction equation t=279.6873-1.9591X1-1.5867X2-12.9721X3-10.5595X4+1.3076X5-7.9399X6-1.7314X7-17.0574X8-17.1047X9-1.9760X10.Based on the TVB-N value of mutton at room temperature,a hyperspectral binding kinetic model was established,and got the prediction equation t=ln(-19.2429+1.9078X1-13.2759X2+0.7971X3+2.4341X4+4.5200X5+4.4762X6+2.5836X7+2.3405X8+5.9684X9+6.4208X10+5.8924X11+5.8309X12+5.5257X13+4.9769X14+4.9613X15+4.8641X16+4.4409 X17+2.2024 X18+2.1198 X19+1.9605X20+1.9801X21)1/0.169.Based on the TVB-N value of mutton at cold and fresh temperature,a hyperspectral binding kinetic model was established,and got the prediction equation t=-2.7774+0.5792 X1+0.7187 X2-1.4045 X3+0.4849 X4-0.9005 X5+1.0779 X6-0.6328 X7+1.1935 X8-1.0432 X9.Therefore,when storing at room temperature,the prediction equation of storage days can be obtained with pH or TVB-N as the key indicators;when storing at cold temperature,the prediction equation of storage days can be obtained with TVB-N as the key indicator. |
PDF Full Text Request