Calibration of visible and near-infrared spectral imaging technology to predict the quality evolution of retail fresh pork bellies with different fat content.

This study investigates quality changes occurred in sliced pork belly with different fat content during refrigerated storage, and the potential of spectral imaging technology in predicting quality properties. Pork bellies with different fat levels (low 'LF', medium 'MF' and high 'HF') were selected from slaughtering houses and directly transferred to the laboratory. The sliced bellies were packed in modified atmosphere packages with high oxygen levels (80 %) and the essential visual and olfactory characteristics, microbiological load, pH, lipid oxidation and colour values were assessed throughout 20 days of refrigerated storage. The spectral images of all belly samples were acquired in the wavelength range from 386 to 1015 nm. Results revealed significant quality losses throughout storage attributed primarily to lipid oxidation and colour changes. The HF bellies showed lower L* and higher a* values than LF and MF. Additionally, the LF and MF bellies, with higher proportion of polyunsaturated fatty acids, showed higher lipid oxidation compared to the HF bellies throughout storage. The appropriate combination of spectral preprocessing, together with the appropriate selection of the region of interest, facilitated the development of robust models to predict the visual appearance, odour, lipid oxidation, and a* values of belly slices during refrigerated storage. The obtained results demonstrated the potential of spectral imaging for predicting quality characteristics of sliced fresh pork bellies during refrigerated storage.