Matching of Scattering Centers Based on 3-D Scattering Center Model with Application to SAR ATR

This letter proposes a scattering center matching method based on 3-D scattering center model with application to synthetic aperture radar (SAR) automatic target recognition (ATR). The 2-D scattering centers predicted by 3-D scattering center model are matched with the extracted scattering centers from the test image using a neighbor matching algorithm. The matching result is quantitatively evaluated by five indexes. The neighbor matching efficiently builds robust correspondences between the model-predicted and extracted scattering center sets. And the five indexes comprehensively evaluate the similarity between the two scattering center sets. Afterwards, a one-against-one voting strategy is performed based on the five indexes for target recognition. To evaluate the effectiveness and robustness of the proposed method, experiments are conducted on the simulated data of three ground targets by electromagnetic (EM) code under various conditions.