A Portable System for Prediction of Pork Freshness Parameters Using Multispectral Imaging Technology

Abstract

<abstract> <bold>Abstract.</bold> Quality of fresh meat is the focus of attention in meat industry at home and abroad. Pork is important source of meat, and its freshness is one of the most important factors that affect its quality. This paper studied on application of multispectral imaging technology to predict pork freshness parameters rapidly and nondestructively. A portable multispectral imaging system was developed, and a Lorentzian distribution (LD) function was used to fit the scattering curve at the selected wavelengths. Next, average value of image grayscale and definite integration of scattering curve were adopted to establish pork freshness parameters prediction model with LD function parameters. In this study, fresh pork total volatile basic nitrogen (TVB-N), pH value and color (L*, a*, b*) were chosen as freshness parameters whose prediction modes were established by multiple linear regression (MLR) method. By comparison, the results of using definite integration of scattering curve combined with LD function parameters to predict pork freshness parameters were best with correlation coefficient (r) of TVB-N prediction model of 0.84 ,and the prediction standard error (SEP) of TVB-N of 2.23mg/100g;the r of pH value was 0.78 and the SEP was 0.05; the r of color parameter L* was 0.90 with SEP 0.73; the r of a* was 0.96 with SEP 0.56; the r of b* was 0.91 with SEP 0.52.The results show that multispectral imaging technology is a promising method to predict pork meat freshness parameters rapidly and nondestructively.