Automatic Target Classification of SAR Images Using Global Scattering Center Model

Abstract:
Model based target classification method has many advantages over template based ones, such as smaller database, smaller computation load for search and match, flexibility in changed operation condition, etc.. This paper introduces a SAR image classification method based on global scattering center model. We propose to extract regional feature from SAR images and match it with the scattering center templates. This region-to-point match scheme is easier to implement and is more robust to non-ideal factors than the point-to-point match scheme. Experiments using data computed by a high frequency electromagnetic code verify the validity of the method.