BE- Uydu Haberleşmesi ve Uzaktan Algılama Lisansüstü Programı - Doktora

Bu koleksiyon için kalıcı URI

http://hdl.handle.net/11527/257

Gözat

Konu "Saçılım" ile BE- Uydu Haberleşmesi ve Uzaktan Algılama Lisansüstü Programı - Doktora'a göz atma
  • Öge
    Direct and inverse electromagnetic wave scatteringrelated to rough surfaces
    (Lisansüstü Eğitim Enstitüsü, 2020) Sefer, Ahmet ; Yapar, Ali ; 657633 ; Uydu Haberleşmesi ve Uzaktan Algılama Bilim Dalı
    Electromagnetic wave scattering from one dimensional rough surface is analyzed both direct and inverse problems point of view. The surface could be a deterministic one or have a random variation which can be characterized by a stationary stochastic random distribution. In the case of direct problem, with the knowledge of the surface roughness and the medium parameters, the field distribution in the whole space is obtained in terms of surface integral representations by means of appropriate integral kernels. The problem is then reduced to an integral equation or a system of an integral equations using suitable boundary conditions. To be more precise, if the problem needs to apply a single boundary condition, then it is expressed by a single integral equation. In case of more than one boundary condition requirements such as rough surfaces forms a boundary between two penetrable media, it is expressed by a system of an integral equations. The solution of the integral equation (or the system of integral equations) for the direct problem is accomplished by a traditional numerical method called method of moments (MoM). This method base on expressing the change of the field -or its normal derivative- distribution on the surface via appropriate basis functions and thus reducing the corresponding integral equation to a matrix equation. In addition to the classical integral equation solution of the problem, a spectral domain integral equation solution based on one-dimensional Fourier transform and Taylor expansion is presented as an alternative and new approach. The accuracy, validity limits and the efficiency of this new approach are analyzed with appropriate comparisons. In the second main part of the thesis, inverse problem algorithms are presented in which the geometry of an inaccessible rough surface is tried to be determined by the scattered field data. Inverse problem algorithms are essentially based on the arrangement of the integral equations used in the solution of the direct problem such that both field -or its derivative- on the surface and the rough surface geometry is included to the equations as the unknowns. Then, this nonlinear integral system is solved iteratively. In this context, a linearization based on Newton method and containing Freechet derivatives was applied in order to determine the surface function, which is the main unknown of the inverse problem. Finally, the ill-posed system was regularized by Tikhonov and solved in the sense of least squares approach. Within the scope of the thesis study, three scenarios for both direct and inverse problems have been considered. Two of these three scenarios cover the analysis of rough surfaces with perfectly electric conducting (PEC). PEC surfaces considered in this context can be illuminated by the TE or TM polarized incident wave. Since the illumination of the surface with each polarization requires consideration of different boundary conditions for the solution, TE and TM cases were analyzed separately. Finally, the scenario where the roughness is positioned to separate the two dielectric media is analyzed. The feasibility of the algorithms has been tested through numerous simulations and the obtained results are discussed in detail.