Reflectance Imaging Spectroscopy (RIS) relies on measuring the radiation reflected from a surface as a function of wavelength. The hyperspectral sensor captures a three‑dimensional volume – the data cube – where the two spatial dimensions (x, y) are combined with a third spectral dimension (λ). Typical spectral resolution ranges from 2 to 10 nm, allowing the detection of subtle absorption differences characteristic of organic and inorganic pigments.
Cube preprocessing includes radiometric correction, noise removal, and normalization against a white reference. Subsequently, reflectance curves are compared with a library of standards using similarity metrics such as the Vector Angle (VA). The VA mapper computes the angle between a pixel spectrum and each reference; the smallest angle indicates the best match, producing pigment distribution maps.
Principal Component Analysis (PCA) reduces cube dimensionality, highlighting dominant spectral variations and facilitating the visualization of patterns invisible to the naked eye. False‑color infrared (IRFC) images exploit near‑infrared bands to reveal organic binders or deeper paint layers.
Complementary techniques – XRF, Raman, FT‑NIR, and FT‑MIR – provide elemental and vibrational information. XRF detects elements like Fe, Cu, Zn, confirming the presence of oxide or metal‑salt pigments. Raman identifies molecular bonds, distinguishing organic pigments (e.g., alizarin) from inorganic ones (e.g., Prussian blue). FT‑NIR is sensitive to O‑H, N‑H, and C‑H vibrations, useful for binders and organic pigments, while FT‑MIR covers the mid‑infrared region, revealing functional groups typical of inorganic compounds.
RIS limitations include surface topography sensitivity, the need for precise calibration, and possible Raman fluorescence interference. Nevertheless, integrating multiple portable techniques enables comprehensive characterization while minimizing impact on the artwork, delivering valuable data for preventive conservation.
