Hyperspectral Imaging for Mycotoxin Monitoring in Food: A Comprehensive Review of Technologies, Algorithms, and Applications

Mycotoxin contamination represents one of the most pressing food safety challenges worldwide, with millions of tons of agricultural commodities affected annually. Conventional detection methods, while accurate, are labor-intensive, destructive, and unsuitable for large-scale screening. Hyperspectral imaging (HSI) has emerged as a transformative non-destructive analytical technique capable of simultaneously capturing spatial and spectral information across hundreds of contiguous wavelengths. This review critically evaluates the current state of HSI technology for mycotoxin detection in food products, covering the physical principles underlying spectral–mycotoxin interactions, systematic applications across major mycotoxin classes (aflatoxins, deoxynivalenol, ochratoxin A, fumonisins, and zearalenone), and the integration of machine learning and deep learning algorithms for spectral data analysis. The review reveals that while Vis-NIR (400–1000 nm) and SWIR (1000–2500 nm) HSI systems have achieved classification accuracies exceeding 90% for several mycotoxin–matrix combinations, fundamental challenges persist in model transferability, direct quantification at regulatory thresholds, and scalability for industrial deployment. Recent advances in transformer architectures, transfer learning, interpretable deep learning, and portable multispectral systems demonstrate encouraging progress toward practical implementation. This review concludes by identifying critical research gaps and proposing strategic directions for translating HSI-based mycotoxin detection from laboratory proof-of-concept to routine industrial application.