Maritime target classification from SLC SAR data Spectral Profiles

Abstract:
In this study, an approach for automatic target recognition in Synthetic aperture radar (SAR) imaging, based on the spectral information of the SAR data is introduced. This work focuses on developing a binary classifier exploiting the spectral information from different ships in the single look complex format (SLC), and the performance of the proposed approach is evaluated on SLC (before and after radiometric calibration) SAR data from Sentinel-1. A Recurrent Neural Network is developed for the classification task and the classes “cargo” and “tanker” are selected for the assessment. As a result, an accuracy of over 69% is obtained in the recognition.