An overview of robust maritime situation awareness methods

Maritime situation awareness is the core of ship navigation and determines the development direction of smart ships. Visual perception becomes an active yet important research task in maritime traffic. To aim that, this paper introduces three situation awareness methods based on visual images. The first method uses kernelized correlation filtering algorithms and Kalman filter models to solve ship occlusion in the image and achieve accurate ship tracking tasks; The second method uses kernelized correlation filtering and logarithmic polar coordinate transformation to solve the imaging size changes during ship navigation; The third method uses yolo algorithms to obtain the high-quality distance information from port videos. The experimental results and quantitative indicators show that each method can effectively achieve reliable maritime situational awareness tasks. The comprehensive realization of maritime situation awareness can better reduce risks and ensure the safety of ships.