Robustness of inverse solutions for radiative transfer parameters from light signals measured with different detection configurations

Abstract

Spectroscopic characterization of turbid samples needs to solve inverse scattering problems (ISPs). With the radiative transfer (RT) theory as the framework for modeling light-matter interaction, accurate characterization becomes feasible but practical implementation remains challenging. We have evaluated the uniqueness and robustness of ISP solutions for a multiparameter spectrophotometer system without integrating spheres using different detection configurations for signal measurement. A novel parameter of area descending rate (ADR) has been developed to sample globally the distribution of an objective function in the 3D space of RT parameters and quantify uniqueness and robustness of ISP solutions. Analysis of the objective function for different sample types show the uniqueness of solutions for different detection configurations with increased complexity for samples of very larger single-scattering albedo values. The robustness of ISP solutions as quantified by ADR was found to be insensitive to the changes in detection configurations. Such stability provides strong evidences that the multiparameter spectrophotometry using only three photodiodes combined with Monte Carlo simulations yields a valuable and practical approach for turbidity characterization.