Use of three tristimulus values from surface reflectance spectra to calculate the principal components for reconstructing these spectra by using only three eigenvectors

Abstract

The division of Color Space into ten zones, corresponding to the ten Munsell hues, allows a good reconstruction of surface reflectance spectra using just three eigenvectors, obtained by applying principal components analysis to the reflectance spectra of the Munsell Atlas specimens (model group), although the basis vectors obtained are different for each subspace. The use of the tristimulus values from each measured spectrum, calculated with the Illuminant D65 and the Standard Observer CIE64 to obtain the principal components necessary to reconstruct the spectrum, allows a very high degree of metamerism to be attained between the two spectra (measured and reconstructed). Furthermore, this method of calculating the principal components allows reconstruction of the spectra of specimens from other sample sets that differ from the model group used in the PCA. The colorimetric accuracy obtained in the new sample sets is similar to that obtained for the model group.