Radar image reconstruction in an inhomogeneous medium using an arbitrary array

The authors have previously (Wakayama et al., 1992) developed an algorithm for radar image reconstruction which can be applied to an arbitrarily arranged antenna array. The algorithm is based on a model fitting by nonlinear least-squares. In the present paper, this algorithm is extended to treat a case where the medium has simple inhomogeneity. Reconstruction is executed on the assumption that targets consist of discrete point scatterers embedded in a medium composed of two layers with different permittivity. The algorithm of the model fitting including the initial guess for the fitting is described. Nonlinear least-squares fits the theoretical waveform calculated by the FDTD method to the observed waveform, and determines the parameters of the targets and the medium. The proposed algorithm estimates the position of targets more accurately than the conventional algorithm which assumes homogeneity of the medium