Study on Technology for Estimating the Distance and Pose of a Ship from a Quay Wall Using Aerial Imagery

This paper focuses on estimating the distance and status of ship during berthing and unberthing using Unmanned Aerial Vehicle (UAV). The goal is to develop a technology that estimates the distance and angle of the ship relative to the quay wall based on aerial images captured by UAV. The research subject is Pukyong National University's research vessel 'Nara,' and the berthing and unberthing process was measured to conduct the study. The aerial images captured by UAV are utilized for measuring the ship and quay wall using the YOLO (You Only Look Once) v8-seg computer vision-based deep learning model. The ship's pose within the image was estimated using Oriented Bounding Box (OBB) with the measured ship's point data. For quay wall line detection, methods based on the RANSAC (RANdom SAmple Consensus) algorithm and one-dimensional quay wall data linearization were compared and analyzed to determine the most effective approach. This process significantly enhances the ship's berthing and unberthing capabilities, supporting safe and efficient maritime transportation. This study provides important insights into ship position and pose estimation during the berthing process and suggests new directions for the development of UAV-assisted ship berthing technology.