A Quasi Linear Sampling Method in Electromagnetic Inverse Scattering

We consider the electromagnetic inverse scattering problem of reconstructing the shape of perfect conducting scatterers, from knowledge of incident plane waves and the far-field patterns of scattered fields. Traditional Linear Sampling Method identifies the edge based on information from a large number of points, sampled from reconstruction region. In this case, not all the sampled points locate at the edge of scatterer. Avoiding sampling these unnecessary points, Quasi Linear Sampling Method is proposed. This new adaptive method determines an approximate initial point 'on' the edge of scatterer through a pre-defined empirical threshold value. From this initial point, other points on the edge are adaptively traced, under the assumption that the edge of scatterer is a continuous curve. Comparison of this newly proposed method with the old one shows that, improvement is achieved in decreasing computational effort while maintaining the similar reconstruction resolution.