SHIP DETECTION AND MULTI-TEMPORAL SHIP MONITORING USING GOCI-II

Abstract

Detection and tracking of maritime ship are an area being studied by all countries and organizations for use in various fields such as accidents, navigation management, and illegal activity surveillance. Recently, the use of satellites such as high-resolution satellite images and SAR (Synthetic Aperture Radar) for ship detection is increasing, and the importance of maritime surveillance devices such as cluster satellites to establish a maritime surveillance system is increasing. Until now, vessel detection has been mainly accomplished using polar orbiting satellites, but the satellite's revisit period is long, which limits continuous ship monitoring. In this study, the possibility of multi-period ship monitoring using the geostationary satellite GOCI-II was confirmed for ship detection and ship movement pattern inference. GOCI-II is a low-resolution image for monitoring the marine environment, but can identify ships using the difference in infrared reflectance between the sea and ships. From GOCI-II images and Sentinel-2 images, it was possible to determine the location of large ships passing through the Korean Peninsula within a range of about 500m and estimate their speed and direction. We verified the ships detected in GOCI-II by collecting the vessel's identity, movement path, and speed through AIS received on land. The results of this study are expected to have a significant impact on future satellite-based ship monitoring technology through linkage with high- resolution satellite images and multi-band calculations