Deep learning-assisted spectral technology monitoring the preservation effect of sodium octenyl succinate starch co-loaded with cinnamaldehyde and carvacrol microcapsules for pork.

Green encapsulation materials prevent meat spoilage and quality deterioration by providing sustainable protection, which plays a crucial role in enhancing meat freshness. Therefore, sodium octenyl succinate starch (OSA starch) co-loaded with cinnamaldehyde (CA) and carvacrol (CAR) (CA/CAR/OSA) microcapsules were prepared by spray drying, using gelatinized OSA starch as the wall material and Tween 80 as the emulsifier in the current study. The interaction between OSA starch and Tween 80 formed a network-like wall structure, stabilizing the embedding of CA and CAR. The microencapsulation embedding efficiency (EE) reached 96.33 ± 0.87%, with a production yield of 63.82 ± 3.0%. Furthermore, the microcapsules exhibited low hygroscopicity, high storage stability, and synergistic antibacterial activity. CA/CAR/OSA was used for pork preservation, with the TVB-N and pH values remaining within the normal range at 14 days, effectively extending the shelf life of the pork. To address the problems of time-consuming and destructive of traditional detection methods, a rapid pork quality detection model was developed by combining visible and near-infrared (Vis-NIR) spectroscopy with deep learning (DL), achieving high predictive accuracy for preservation days (R2 = 0.9862). This study combining CA/CAR/OSA with DL provided a highly accurate prediction model for pork freshness based on lab-acquired spectral data.