Buried object adaptive shape reconstruction and ground parameters estimation using differential evolution

In this study, the capability of the opposition-based differential evolution stochastic searching algorithm in shape reconstruction of a two-dimensional conducting target buried in a lossy dielectric half-space is investigated using simulated backscattered fields calculated above the ground by both method of moment with far-field approximation and geometrical optics approximation at different observation and frequency points. For an efficient reconstruction, the target boundary is approximated by a non-uniform cubic B-spline curve with variable number of curve control parameters (CP). Using the knot insertion Oslo algorithm and through the optimisation process, the DE algorithm finds the optimum number of CP, successfully reconstructs the ground parameters (i.e. permittivity and conductivity) together with target upper part shape and location for moderate values of the signal-to-noise power ratio.