Sparse electromagnetic imaging

We exploit the sparseness of the induced currents on the surfaces of perfectly conducting (PEC) scatterers to develop a non-iterative imaging algorithm in two-dimensional (2D) space. We assume that the targets can be suitable represented by a set of unknown filament currents as in the equivalent source method. We apply the surface equivalence theorem and l₁ regularization on the model with the filament currents. In this way, we promote the solutions with small number of equivalent sources placed in the vicinity of scatterer surfaces. We show that the algorithm reconstructs reliably the shape of the hidden scatters using single frequency data.