Civilian vessel classification with COSMO-SkyMed images based on feature analysis

Vessel classification is an important part of global monitoring of maritime traffic and ocean environment. In high resolution SAR images, it is possible to recognize civilian vessels base on their features of back scattering and structure. This paper presents a vessel classification method that uses single-pol COSMO-SkyMed images to categorize vessel targets into three types, which include cargo vessels, container vessels and oil tankers. The result shows that this method can classify vessels at an average precision of 80%.