Linear modelling is a common method to efficiently represent object reflectance spectra. In this paper, an alternative method is used based on the physical characteristics of the spectral reflectance. Because of the band-limited properties of the spectral reflectance and the smoothness of the spectral reflectance curves, we model the reflectance spectra by using modified Gaussian functions to represent absorption properties of the objects, which are physically the primary cause of the spectral reflectance. A set of artificial spectral reflectance generated from this physical model showed very similar characteristics of the Munsell data set, and the reverse of the model can also give accurately reconstruction of the original reflectance spectra.
[1]
L. Maloney.
Evaluation of linear models of surface spectral reflectance with small numbers of parameters.
,
1986,
Journal of the Optical Society of America. A, Optics and image science.
[2]
J. Parkkinen,et al.
Characteristic spectra of Munsell colors
,
1989
.
[3]
Roy S. Berns,et al.
A review of principal component analysis and its applications to color technology
,
2005
.
[4]
K. Nassau.
The fifteen causes of color: The physics and chemistry of color
,
1987
.
[5]
Carle M. Pieters,et al.
Deconvolution of mineral absorption bands: An improved approach
,
1990
.