Estimating the best illuminant for art paintings by computing chromatic diversity

Abstract

Hyperspectral imaging over the visible range (400 - 720 nm) in 10 nm steps was used to digitalize oil painting on wood in the Museu Nogueira da Silva, Braga, Portugual. The hyperspectral imaging system (HIS) used to capture the images was developed by Nascimento et al (2002 Journal of the Optical Society of America A--Optics Image Science and Vision 19 1484 - 1490) and improved by Foster et al (2004 Visual Neuroscience 21 331-336). The HIS was composed by a low-noise Peltier-cooled digital camera with a spatial resolution of 1344 × 1024 pixels (Hamamatsu, C4742-95 - 12ER), and a fast - tuneable liquid - crystal filter (VariSpec, model VS - VIS2 - 10HC - 35 - SQ, Cambridge Research & Instrumentation, Inc., MA, USA) mounted in front of the lens, together with an infrared blocking filter. The spectral reflectances of each pixel of the paintings were estimated from a grey reference surface present in the scene. Illuminant spatial non -­‐ uniformity was corrected by a white paper reference measured prior to painting image acquisition. The radiance from each painting under D65, A, Solux, habitual museum illumination (halogen and xenon lamps) and fluorescent lamps of correlated colour temperature (cct) 2940, 4230 and 6500 K was estimated and the corresponding luminance and chromaticity distributions computed. In each case, the number of discernible colours was estimated by computing the painting representation in CIELAB space and by counting the number of non - empty unit cubes in that space (Pointer et al, Color Research and Application 23 52 - 54; Linhares et al, 2004 Perception 33 65 – 65)