Application of portable visible and near-infrared spectroscopy for rapid detection of cooking loss rate in pork: Comparing spectra from frozen and thawed pork

In this study, a portable visible and near-infrared (Vis-NIR) spectroscopy system was developed to quickly assess the cooking loss rate in pork, with findings comparing spectra from frozen and thawed pork. Firstly, the Vis-NIR spectral data of samples were collected under freezing and thawing conditions. After selecting characteristic variables using four different variable selection algorithms, partial least square (PLS) was used to predict cooking loss rate. The competitive adaptive reweighted sampling PLS (CARS-PLS) models were noted with higher prediction results, based on correlation coefficients of calibration (R-c) = 0.8362 and prediction (R-p) = 0.8154 for frozen pork samples spectra, while R-c and R-p for thawed pork samples spectra were noted as 0.8748 and 0.8421, respectively. The results of the comparison showed that the prediction effects of frozen and thawed pork spectra were similar. The current method has an excellent prospect to predict frozen pork quality without thawing.

Automated summary

A portable visible and near-infrared (Vis-NIR) spectroscopy system was developed to assess pork cooking loss rate, comparing spectra from frozen and thawed pork. Partial least square (PLS) was used to predict cooking loss rate after selecting characteristic variables using four different algorithms. The results showed that the competitive adaptive reweighted sampling PLS (CARS-PLS) models had higher prediction results for both frozen and thawed pork spectra. The method has the potential to predict frozen pork quality without thawing.