Adaptive Selective Kernel Prototype Learning for Imbalanced SSE
摘要
The vitality of the world’s oceans is diminishing, jeopardizing the underpinning of the international marine economy. In response to heightened environmental awareness, the maritime shipping sector is under significant pressure to curtail carbon output while ensuring operational cost-efficiency. More demanding maritime regulations are accelerating the adoption of autonomous marine systems, viewed as crucial for advancing sustainability agendas. Nevertheless, the optimal functioning and decision-making capabilities of these automated vessels are critically reliant upon the availability of precise, real-time sea condition assessments [1, 2]. Sea state refers to the conditions of the sea surface concerning waves and wind at a particular location and moment, representing a constantly changing environment driven by wind patterns and duration [3]. Essential descriptors include wave height, wind speed and direction, wave frequency, and energy distribution.