Time Domain Image Reconstruction for Homogenous Dielectric Objects by Dynamic Differential Evolution

• Published in: Electromagnetics Vol. 30; no. 4; pp. 309 - 323
• Main Authors: Sun, Chi-Hsien, Chiu, Chien-Ching, Li, Ching-Lieh, Huang, Chung-Hsin
• Format: Journal Article
• Language: English
• Published: Philadelphia Taylor & Francis Group 01.05.2010; Taylor & Francis Ltd
• Subjects: Cylinders; Dielectric properties; dynamic differential evolution; Dynamics; Electric fields; Electromagnetics; Evolution; finite difference time domain; Finite difference time domain method; homogenous dielectric cylinder; Image reconstruction; Mathematical models; Normal distribution; Numerical analysis; Reconstruction; Studies; sub-gridding technique; Time domain; time domain inverse scattering
• ISSN: 0272-6343; 1532-527X
• DOI: 10.1080/02726341003712491

This article presents the studies of the microwave image reconstruction of two-dimensional homogeneous dielectric cylinders that are based on the finite difference time domain method and dynamic differential evolution. For forward scattering, the finite difference time domain method is employed to calculate the scattered E-fields; for inverse scattering, the dynamic differential evolution scheme is utilized to determine the shape, location, and permittivity of the cylindrical scatterers with an arbitrary cross-section. The subgirdding technique is implemented for the finite difference time domain code in order to more smoothly model the shape of the cylinder. In addition, in order to more effectively describe an unknown cylinder with an arbitrary cross-section during the course of searching, the closed cubic-spline expansion is adopted to represent the scatterer contour instead of the frequently used trigonometric series. Numerical results demonstrate that, even when the initial guess is far from exact, good reconstruction can be obtained. In addition, the effects of Gaussian noise on the reconstruction results are investigated. Numerical results show that even the measured scattered fields are contaminated with Gaussian noise, and dynamic differential evolution is able to yield good reconstructed quality.