Abstract
Shape reconstruction from scattered electromagnetic fields is considered for two-dimensional through-wall imaging (TWI). The finite-difference time-domain method is applied to calculate the scattered E fields for forward scattering. Inverse scattering problems are transformed into optimisation problems based on the scattered E fields. A dynamic differential evolution (DDE) stochastic searching algorithm for shape reconstruction of a two-dimensional conducting target hidden behind a homogeneous building wall is demonstrated using simulated backscattered fields. The DDE is a population-based optimisation approach that aims to minimise the objective function between mimic measurements and computer-simulated data. Thus, the shape of a metallic cylinder can be obtained by minimising the objective function. Simulations show that DDE can successfully reconstruct TWI for a metallic cylinder. In addition, the effect of Gaussian noise on the reconstruction is investigated.