FDTD-BASED 2-DIMENSIONAL FREQUENCY-DEPENDENT NUMERICAL IMAGING FOR RADAR SCATTERING

This Letter presents FDTD-based two-dimensional numerical frequency-de pendent imaging for radar scattering. Instead of measurement, the imag es are reconstricted from the near scattering field with the use of th e FDTD method. The maximum-likelihood (ML) algorithm is used to yield high-resolution imaging for wideband data. To obtain the initial estim ates of ML algorithm matrix enhancement and the TLS-ESPRIT technique a re repeatedly employed to provide the positions and amplitudes of scat tering centers for each segment narrowband data. The initialization of the discrete geometry parameters is obtained by an approximation base d on the amplitude variations at different frequency segments. Compute r simulation shows the validity of this technique. (C) 1996 John Wiley & Sons, Inc.