Background
Meat and poultry are among the most highly perishable food commodities, and their rapid deterioration presents significant challenges to food safety, quality assurance, and economic sustainability. Conventional spoilage detection methods, including microbiological and chemical assays, remain reliable; however, they are often destructive, time-consuming, and unsuitable for real-time monitoring applications.
Scope and approach
Hyperspectral imaging (HSI) has become a prospective non-invasive technology that could be used in the simultaneous spatial and spectral data collection of the meat surface, offering in-depth information on chemical and microbial transformations. The present review is based on the current use of HSI in the process of monitoring microbial activity, identifying chemical indicators, including total volatile basic nitrogen (TVB-N), Thiobarbituric Acid Reactive Substances (TBARS), and pH changes, and measuring visible changes in colour and texture. Special attention is given to the combination of HSI and chemometric methods and state-of-the-art machine learning models, such as principal component analysis (PCA), partial least squares discriminant analysis (PLS-DA), convolutional neural networks (CNNs), long short-term memory networks (LSTMs), and autoencoders.
Key findings and conclusion
HSI has proven to be very accurate in the early detection of spoilage, in most instances better than conventional means. New innovations, including creation of portable HSI units, artificial intelligence, and multimedia systems that combine HSI and other sensors or internet of things, have expanded its application potential. Nevertheless, there are still notable issues, such as managing massive amounts of data, high prices of equipment, and the lack of uniform protocols to implement in industries. Future studies must aim at minimizing the size of the devices, creation of strong calibration systems, and the ease of incorporation into smart food devices. Together, these developments make HSI a future-proof, non-destructive, and dependable technology to assess the quality and safety of modern meat. This review further demonstrates that hybrid VIS-NIR-SWIR configurations combined with interpretable chemometric and machine-learning models consistently outperform single-range systems, while the lack of standardized validation and regulatory frameworks remains the primary barrier to large-scale industrial adoption.


扫码关注我们
求助内容:
应助结果提醒方式:

