Full Wave 2D Modeling of Scattering and Inverse Scattering for Layered Rough Surfaces with Buried Objects.

Abstract

Efficient and accurate modeling of electromagnetic scattering from layered rough surfaces with buried objects finds applications ranging from detection of landmines to remote sensing of subsurface soil moisture. In this dissertation, the formulation of a hybrid numerical/analytical solution to electromagnetic scattering from layered rough surfaces is first developed. The solution to scattering from each rough interface is sought independently based on the extended boundary condition method (EBCM), where the scattered fields of each rough interface are expressed as a summation of plane waves and then cast into reflection/transmission matrices. To account for interactions between multiple rough boundaries, the scattering matrix method (SMM) is applied to recursively cascade reflection and transmission matrices of each rough interface and obtain the composite reflection matrix from the overall scattering medium. The validation of this method against the Method of Moments (MoM) and Small Perturbation Method (SPM) will be addressed and the numerical results which investigate the potential of low frequency radar systems in estimating deep soil moisture will be presented. Computational efficiency of the proposed method is also addressed. In order to demonstrate the capability of this method in modeling coherent multiple scattering phenomena, the proposed method has been employed to analyze backscattering enhancement and satellite peaks due to surface plasmon waves from layered rough surfaces. Numerical results which show the appearance of enhanced backscattered peaks and satellite peaks are presented. Following the development of the EBCM/SMM technique, a technique which incorporates a buried object in layered rough surfaces is proposed by employing the T-matrix method and the cylindrical-to-spatial harmonics transformation. Validation and numerical results are provided. Finally, a multi-frequency polarimetric inversion algorithm for the retrieval of subsurface soil properties using VHF/UHF band radar measurements is developed. The top soil dielectric constant is first determined using an L-band inversion algorithm. For the retrieval of subsurface properties, a time-domain inversion technique is employed together with a parameter optimization for the pulse shape of time delay echoes from VHF/UHF band radar observations. Some numerical studies to investigate the accuracy of the proposed inversion technique in presence of errors are shown.