Abstract

Preservation of spectral color information of a set of surface colors under 36 different light sources are evaluated. Two measures of information are used, the energy and relative entropy. The energy is assessed by the eigendecomposition method. It is found that while 9 eigenvectors are necessary for full presentation of the spectral reflectances of surface colors, 5–8 eigenvectors are adequate to fully characterize the total variances of radiance spectrum, depending on light sources. Results also show that, the derived eigenvectors are sparse in some wavelengths mostly for the sources with limited numbers of emission bands. It is found that a significant spectral information loss occurs due to the illumination of surface colors. The Kullback- Leibler divergence measure is used to compare the relative entropy between reflectance spectra of objects and the corresponding radiance spectra of samples under each light source. Results show that there are big differences between the light sources in relative entropy between the reflectance spectra and the radiance spectra of objects. Based on energy and relative entropy considerations, the tri-band mixed-color white light LEDs generally have the highest loss of spectral information before the projection of radiance data over the reciver.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call