Study on quality evaluation of the entire stir-frying process of charred stir-fried malt based on intelligent sensing and machine learning

This study aimed to establish objective indicators for assessing the stir-frying degree of charred stir-fried malt (CSFM) to overcome the limitations of traditional empirical judgment. By integrating visual recognition, texture analysis, Heracles NEO electronic nose, and high-performance liquid chromatography (HPLC), we systematically monitored dynamic changes in color, texture, odor, and active components during stir-frying. Results indicated that color parameters (R, G, B) and hardness decreased significantly, with defined quantitative ranges established for CSFM. Odor profiles altered markedly, and five key processing markers, including 5-methylfurfural, were identified. The content of hordenine decreased by 56.6%, while 5-HMF increased strongly and correlated closely with stir-frying time. Other components such as catechin and ferulic acid showed varied trends. The study successfully quantified key attributes of CSFM, identified reliable processing markers, and revealed correlations among multi-dimensional indices, thereby providing a scientific foundation for standardizing the production of CSFM decoction pieces.