무인수상선의 적외선 카메라를 이용한 선박 자동 탐지 기법

Abstract

This study presents the infrared vision-based automatic ship detection algorithm for autonomous navigation of unmanned surface vehicles (USVs). In order to minimize inherent noises that arise in a marine environment, image filtering techniques are applied and the vertical edge and the diagonal edge images divided through 2-D discrete wavelet transform technique are combined without the horizontal edge image. In order to get the wavelet energy map image that clearly shows thetarget ship, an weight function with Gaussian distribution is then applied in the combined image. The feasibility of the proposed approach was validated through field experiment data obtained in a real-sea environment.