Purchase this article with an account.
João Manuel Maciel Linhares, Paulo Daniel Araújo Pinto, Raphael Gonçalves Lopes, Sérgio Miguel Cardoso Nascimento; The number of discernible colors in natural scenes rendered with daylight illuminants for normal and color deficient observers. Journal of Vision 2009;9(14):63. doi: https://doi.org/10.1167/9.14.63.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The spectral composition of daylight varies along the day with significant repercussions in the number of discernible colors that can be perceived in natural scenes. The chromatic content of 50 hyperspectral images of natural scenes, 22 rural and 28 urban, was studied as a function of the correlated color temperature of illuminant D for normal, anomalous and dichromatic observers. The hyperspectral system used a low-noise cooled digital camera with a spatial resolution of 1024 (H) x 1344 (V) pixels, a fast tunable filter and an infrared filter in front of the lens. Images were acquired from 400 to 720 nm in 10 nm steps. The radiance spectrum for each image pixel was converted into reflectance using the scene illuminant recorded at the time of acquisition. For D illuminants with CIE x chromaticity coordinate in the range 0.25 to 0.4 in 0.0075 steps, the CIELAB color volume was computed for each scene. For normal observers, the number of discernible colors was estimated by segmenting the resulting CIELAB color volume into unitary cubes and by counting the number of non-empty unitary cubes. It was assumed that all the colors that were represented inside the same unitary cube could not be discernible. For color deficient observers, the number of colors was estimated using models simulating for normal trichromatic observers the appearance for color deficient observers. It was found that the number of discernible colors increased with the x chromaticity coordinate of the illuminant for all classes of observers, except for protanopes. The increment was on average of about 13% and of about 14% and 11% for urban and rural images, respectively. It was about 48%, 39%, 3% and 5% for deuteroanomalous, protoanomalous, tritanopes and deuteranopes observers respectively; for protanopes the decrease was of about 7%. These results suggest that illuminants with higher x chromaticity or low correlated color temperature may be the best for maximizing the chromatic content of natural scenes for all kind of observers, except for protanopes.
This PDF is available to Subscribers Only