In autonomous driving technologies, a camera is necessary for establishing a path and detecting an object. Object recognition based on images from several cameras is required to detect impediments in autonomous ships. Furthermore, in order to avoid ship collisions, it is important to follow the movements of recognized ships. In this paper, we use the Singapore Maritime Dataset (SMD) and crawling image for model training. Then, we present four YOLO-based object recognition models and evaluate their performance in the maritime environment. Then, we propose a tracking algorithm to track the identified objects. Specially, in evaluation with high-motion video, the proposed tracking algorithm outperforms deep simple online and real-time tracking (DeepSORT) in terms of object tracking accuracy.
Recently, an offshore support vessel is being widely used to install an offshore structure such as a subsea equipment which is laid on its deck. The lifting operation which is one of the installation operations includes lifting off, lifting in the air, splash zone crossing, deep submerging, and finally landing of the structure with an offshore support vessel crane. There are some major considerations during this operation. Especially, when lifting off the structure, if operating conditions such as ocean environmental loads and hoisting (or lowering) speed are bad, the excess of tension of wire ropes of the crane and the collision between the offshore support vessel and the structure can be occurred due to the relative motion between them. To solve this problem, this study performs the lifting simulation while the offshore support vessel installs the structure. The simulation includes the calculation of dynamic responses of the offshore support vessel and the equipment, including the wire tension and the collision detection. To check the applicability of the simulation, it is applied to some lifting steps by varying operating conditions. As a result, it is confirmed that the conditions affect the operability of those steps.
1 October 2017 in Advances in Engineering Software