Multistatic Uniform Diffraction Tomography Derived Structural-Prior in Bayesian Inversion Framework for Microwave Tomography

In this work, a quantitative Bayesian inversion framework for microwave tomography (MWT) is coupled with a multistatic uniform diffraction tomography (MUDT) method to improve the imaging quality. The method is applied for an industrial use-case of MWT in which we estimate the 2-D spatial distribution of moisture (in terms of dielectric constant) in a polymer foam. In essence, we modify the prior information in the single-frequency Bayesian inversion framework using high-resolution complementary structural information of the imaging domain from a qualitative approach MUDT utilizing broadband frequency-domain data. This way of obtaining structural prior information is effective as it utilizes the data from the same microwave sensor setup in contrast to the frequency-hopping approach, priors derived for other imaging modalities or radar-based techniques with the co-located sensor using, for example, uniform diffraction tomography (UDT) inversion framework. Proposed algorithm performance is tested for different moisture scenarios in the polymer foam with 3D numerical and experimental data from our developed MWT system. It is shown that the proposed approach significantly improves the reconstruction accuracy for the considered cases over just using the Bayesian inversion approach.