Spatial Mismatch Correction Based on 3D Electromagnetic Scattering Model

3D electromagnetic scattering models(EM-model) can predict features at different angles for matching, which plays an important role in the automation target recognition for synthetic aperture radar imagery(SAR-ATR) system. However, there is a space mismatch between the model and the measured image. Therefore, the EM-model and the measured image are in different coordinate systems. To solve this problem, this paper proposes a new spatial mismatch correction algorithm based on EM-model. Considering that the model can be characterized by the position and amplitude of the scattering center(SC), we take the scattering center as the main feature. The scattering centers of the model and the measured data are extracted separately, and the Hausdorff distance(HD) is utilized to measure the distance between EM-model and the measured image. According to the rigid transformation relationship between the EM-model and the measured image, the corresponding rigid transformation model is established. Experimental results based on the moving and stationary target acquisition and recognition (MSTAR) dataset demonstrate the validity of the proposed algorithm.