A stochastic tomographic technique for microwave NDE

This paper presents a stochastic optimization approach to microwave tomographic imaging for nondestructive evaluations (NDE). In particular, a reconstruction procedure is developed in the spatial domain starting from the integral equations of the inverse scattering. A multi-illumination multiview imaging configuration is considered, with the assumption that measurements can be performed around the object under test. Cylindrical dielectric objects of arbitrary cross sections are reconstructed under transverse magnetic illumination conditions. After discretization of the continuous model, the problem solution is recast as a nonlinear optimization problem. A fitness function is defined and minimized by applying a genetic algorithm.