FORMALISATION OF THE SURFACE OBJECTS IDENTIFICATION PROCESS IN AUTONOMOUS SHIP CONTROL BASED ON NEURAL NETWORK MODEL

Abstract

The aim of the study is to develop approaches that ensure automated identification of surface marine objects in real time by formalising the identification process based on a neural network model. The aim of the article is achieved by solving the problem of automated identification of marine objects based on convolutional neural networks. The author proposes an identification method that includes determination of identification parameters aimed at solving the problems of navigation and ensuring safety of navigation. As part of the identification process, the marine objects were classified according to their importance and potential danger to navigation. The study found that a convolutional neural network is the best option for solving the task of identifying surface marine objects. The use of such a neural network allows for efficient image processing and automatic extraction of features that help in identification. This approach can greatly facilitate the work of navigators and increase the level of safety at sea. As a result, the implemented identification algorithm, parameterisation, and use of a convolutional neural network allowed us to create a powerful toolkit for identifying marine surface objects. The implementation of this approach can help improve maritime safety and the efficiency of ship operators in various maritime situations. The identification method is based on convolutional neural networks (CNN). The identification method was tested on a dataset of images of marine objects. The dataset included images of various types of marine objects such as ships, fishing boats, lighthouses, and others. The identification method was able to correctly identify most of the objects in the images, even in difficult conditions such as low light or fog. This method for identifying marine objects can make a significant contribution to improving maritime safety. The method can help navigators avoid collisions with other ships and other objects at sea. The method can also help navigators make better decisions in complex maritime situations.