Identification of Ship's Waterline via the Cooperation of Autonomous Underwater Vehicles

The waterline of ship is very important to navigate safely. However, due to the dynamic and complex marine environment, it is difficult to identify the waterline. This paper presents a method of waterline identification via the cooperation of autonomous underwater vehicles (AUVs). Specifically, we first construct a waterline monitoring system, which consists of shore-based cameras, AUVs, deployment unit and a shore-based control center. Based on this, the waterlines on the side of the ship facing the shore are captured by the shore-based cameras, while the waterlines on the opposite side are captured by AUVs. Using the collected images, we design a waterline recognition algorithm based on canny operator and YOLOv5 network. To recognize the ship's attitude, we propose a frequency-domain-based waterline value calculation algorithm. Finally, experiments verifies the method we proposed is effective.