KN.TOOSI university of technology
This study describes a hybrid technique for identification of buried objects. The objects shape and electromagnetic profile are reconstructed from evaluations of electrical permittivity and conductivity. The method suggests a combination of linear sampling and optimization. Linear sampling method (LSM) is used to recover shape and metaheuristic optimizations essential to reconstruct the inside profile. A special case of a non-homogenous object is considered. In this case the object is a buried homogenous dielectric for which the position and volume of the hole inside is not known. In this case, after recovering data by the same process used in the first step, it is necessary to recover the volume and position of the hole. This case will be useful when the final target of identification is an embedded object. The final results of this case verify the effectiveness and feasibility of the proposed method. Effectiveness of the method is validated by application of one real world cases involving civil engineering. In all cases presented in this study, the rates of final reconstruction error are acceptable and less than 10%. These results prove that this proposed procedure can be applied to reconstruct hollow objects correctly.