Abstract
Ports often experience congestion due to the presence of other ships. This congestion poses challenges for ship berth-ing, and even minor errors during the berthing process can lead to significant accidents. Therefore, it is crucial to accu-rately detect the port surroundings and generate a route for automatic berthing based on this accurate detection. To ad-dress these concerns, this study proposes a method that utilizes LIght Detection And Ranging (LIDAR) to detect nearby objects, such as other ships and quays, and generate automatic berthing routes. LIDAR is known for its high detection accuracy compared to other sensors. However, LIDAR data is typically represented as a point cloud, requir-ing the de-tection of individual objects within it. This study employed a deep learning-based method for precise object detection in the point cloud.