Object recognition based on feature matching of scattering centers in SAR imagery

Extracting high discriminable features has been one of the focuses in SAR ATR research. In this paper, an ATR method is proposed based on sequential feature matching of scattering centers. The feature is extracted from point sets of scattering centers, and estimated by utilizing attributed scattering center model. Then sequential feature matching is employed to determine the identity consistency of vehicle objects. Experiment results of MSTAR data with different azimuth angles and depression angles illustrate good performances.