Hyperspectral imaging for material identification in cultural heritage: a critical review

Hyperspectral imaging (HSI) has emerged as a powerful non-invasive tool for material identification in cultural heritage, integrating imaging and spectroscopy to acquire spatially and spectrally resolved data. This review critically examines the capabilities and limitations of HSI either as a standalone method or in combination with other analytical techniques for identifying materials such as pigments, inks, and dyes in artworks and historical documents. It explores the fundamentals of HSI and recent advances in data analysis, including multivariate analysis, spectral matching, machine learning, and spectral unmixing. Despite these advances, the widespread implementation of this technique in cultural heritage institutions remains limited due to its cost, data complexity, and lack of standardized spectral databases. Future directions include the development of generalized, interpretable identification algorithms and multi-technique data fusion. This review provides conservation scientists and researchers in the field with valuable insights into the realistic potential of HSI for reliable material identification.