A new solution of real-time electromagnetic inverse-scattering

In order to deal with the inverse scattering of dielectric circular cylinder, a new online methodology is applied to the solution of microwave inverse scattering problems. The inverse scattering problem is recast into a regression estimation one by means of a support vector machine(SVM). The approach can be applied to various inverse scattering problems, and it is very suitable to deal with the geometric and dielectric characterization reconstruction and the buried object detection. For the first time this paper applies the support vector machine to the field. The scattered electric field is measured at some points. The SVM is trained using different sample data extracted from the scattered field. The inverse scattering model of dielectric circular cylinder has been setup. The relative dielectric permittivity and the electric conductivity of dielectric circular cylinder are reconstructed and the center position of buried object are detected. The results show that this method is effective and efficient and provides a highly efficient solution for the real-time inverse scattering of objects.