적외선 영상기반 해상물체 탐지 및 추적

Abstract

This study addresses the trajectory estimation of target surface ships using infrared images with the aid of radar for autonomous navigation of unmanned surface ships (USVs). Trajectory estimation using an infrared camera often suffers from low observability since monocular vision has a limited field of view and it basically provides only bearing measurements to the target in the images. Radar is a long-range sensor with a 360-degree field of view, however it has blind zonesat close ranges due to its sensing characteristics. To overcome the shortcomings of individual sensors, an integration method for combining infrared images and radar measurements is implemented and applied to the online trajectory estimation of target ships. The feasibility and performance of the proposed tracking approach are validated through filed experiments with a USV developed by KRISO.