An on-line detection method of pork quality using hyperspectral imaging combined with machine learning

In view of the problems such as serious damage, low prediction rate and poor effect of pork quality, the paper integrates the radial basis function neural network and hyperspectral technology, and establishes the prediction model of pork freshness and water content. The outcomes found that the correlation coefficient of pork freshness correction set and prediction set modeling is 0.9983 and 0.9980, respectively, and the prediction effect is better than other comparison models. The correlation coefficients of pork water content correction set and prediction set modeling were 0.9980 and 0.9986, respectively, and the prediction effect was better than the other comparison models. The outcomes denote that RBF is the best prediction effect of pork quality, which lays a foundation for the online rapid nondestructive detection of pork.