Vol. 42, Issue 4, pp. 737-742 (2012)

Abstract

The principal component analysis is an important method in reflectance spectra reconstruction. This is a research into the effects of a wavelength interval on the performance of reflectance spectra estimation by the principal component analysis technique. The spectral reflectance is reconstructed by using 20, 10 and 5 nm intervals. The RMS accuracy of reflectance spectra fitting is 0.0079, 0.0074 and 0.0072, respectively for 20, 10 and 5 nm intervals with 6 principal components. As the obtained results indicated, the performance of reflectance estimation increased with decreasing the wavelength interval.