Marine Targets Recognition Through Micro-Motion Estimation from SAR data

The capability to perform Automatic Target Recognition (ATR) from SAR images has great importance for both civilian and military applications. However, this task becomes challenging when the quality and quantity of target information is not sufficient to reliably discriminate the targets. This is particularly important when dealing with marine targets, where features such as scattering intensities and shapes are common to many different targets. This paper investigates the possibility to enhance classification capabilities of marine targets in SAR images by exploiting the micro-motion information. This characterizing source of information, is extracted by applying Doppler sub-apertures and pixel tracking on SAR images containing the target of interest. The proposed approach is validated on real COSMO-SkyMed SAR data demonstrating the effectiveness to discriminate ships through their unique Doppler fingerprint.