Ship detection based on feature confidence for high resolution SAR images

Abstract

Ship detection is an important application of global monitoring of ocean environment and maritime traffic. Synthetic aperture radar (SAR) systems are active sensors offering unique good spatial resolution regardless of weather or other conditions. It has been widely used for ship detection. An improved ship detection for high resolution SAR images based on the ship feature confidence is proposed in this paper. The features include kernel density estimation, length-width ratio and the number of target pixels. Targets with high feature confidence will be interpreted as ships. The COSMO-SkyMed SAR image is adopted for investigating the proposed algorithm. Experiment results illustrate that the method can achieve good performances.