MEAN WIND WAVE PERIOD ESTIMATION FROM INDIVIDUAL SHIP NAVIGATION RADAR IMAGES USING DEEP LEARNING METHODS

This paper presents a method for approximating the wind wave period based on ship navigation radar data in a deep learning approach. In this work, a convolutional neural network ResNet is applied, which processes radar images and approximates the values of the wind wave period. The study evaluates the errors of the method in the measure of the standard deviation of the neural network compared to the values measured by the Spotter wave buoy, which allows us to assess the quality of the proposed algorithm. The results demonstrate the efficiency of using neural networks to approximate wave characteristics from radar image data.