ABSTRACT
Time-resolved spectroscopy is a powerful technique permitting the separation of the scattering properties from the chemical absorption properties of a sample. The reduced scattering coefficient and the absorption coefficient are usually obtained by fitting diffusion or Monte Carlo models to the measured data using numerical optimization techniques. However, these methods do not take the spectral dimension of the data into account during the evaluation procedure, but evaluate each wavelength separately. A procedure involving multivariate methods may seem more appealing for people used to handling conventional near-infrared data. In this study we present a new method for processing TRS spectra in order to compute the absorption and reduced scattering coefficients. This approach, MADSTRESS, is based on linear regression and a two-dimensional (2D) interpolation procedure. The method has allowed us to calculate absorption and scattering coefficients of apples and fructose powder. The accuracy of the method was good enough to provide the identification of fructose absorption peaks in apple absorption spectra and the construction of a calibration model predicting the sugar content of apples.
