An electromagnetic model based scattering center detection method by hypothesis testing

This paper presents an approach for synthetic aperture radar (SAR) scattering center detection based on electromagnetic model (EM-model) by hypothesis testing (HT). The scattering center extraction is a critical step in model-based SAR automatic target recognition. In traditional methods, the extraction of scattering centers merely make use of the measured SAR image and lack for the prior information of the target. As a result, the extraction process may be stochastic and unstable which may cause false or missing scattering centers. EM-model can be established offline and predict the scattering center features at different target poses conveniently. In this paper, the target's EM-model is employed to provide prior information for the detection of scattering centers. Then, scattering centers in the measured image is detected one by one through a hypothesis testing approach using the estimated parameters and the model projected parameters. Finally, simulated data predicted by a high frequency electromagnetic software demonstrate the validity of the method.