Abstract

A factor analysis of a set of spectra of multicomponent mixtures yields <i>n</i> orthogonal spectra where <i>n</i> is the number of components in the mixtures. Combinations of these orthogonal spectra, done so as to eliminate the contributions to the spectra from one of the components, yield a set of <i>n</i> – 1 linearly independent abstract spectra. Repetition of this process is used to estimate the spectrum of each of the components without identifying the range of the <i>n</i>-dimensional vector space in which the solution must fall. Several techniques for estimating the appropriate combinations of orthogonal spectra are compared.

Constrained nonlinear method for estimating component spectra from multicomponent mixtures

Keiji Sasaki, Satoshi Kawata, and Shigeo Minami
Appl. Opt. 22(22) 3599-3603 (1983)

Estimation of component spectral curves from unknown mixture spectra

Keiji Sasaki, Satoshi Kawata, and Shigeo Minami
Appl. Opt. 23(12) 1955-1959 (1984)

Component spectra extraction from terahertz measurements of unknown mixtures

Xian Li, D. B. Hou, P. J. Huang, J. H. Cai, and G. X. Zhang
Appl. Opt. 54(30) 8925-8934 (2015)

Identification of pure component spectra by independent component analysis in glucose prediction based on mid-infrared spectroscopy

Sangjoon Hahn and Gilwon Yoon
Appl. Opt. 45(32) 8374-8380 (2006)

Orthogonal Color-Mixture Functions*

D. L. MacAdam
J. Opt. Soc. Am. 44(9) 713-724 (1954)