Method of the horizon detection from electro-optical sensors under maritime environment

Abstract

The article substantiates the relevance of research related to the detection of the sea horizon line in complex environments. Widespread use of CCTV systems is an important step in the detection of moving surface objects over long distances, which will increase the safety of seaports. The analysis of existing methods of detecting the horizon line is carried out. A new method of detecting the sea horizon line is proposed, which allows achieving the goal of reducing the computational cost of searching for targets, and reducing noise from interference caused by changes in lighting from sea waves or clouds. The proposed method consists of a new gradient diagram of gray sections, a calculation of contrast on the basis of a texture, extraction of the field of candidates for the horizon, identification of the point of the candidate and linear fit. In the first stage, the original image is transformed into an image in shades of gray. In the second stage, the threshold values of adaptive segmentation are obtained with the help of an algorithm that creates a set of points of candidates for the sea sky. Finally, a simple clustering method is used to select the corresponding points and their transformation. Experimental results show that the proposed method can accurately determine the line of the sea horizon in difficult cloudy conditions or ocean waves, in the presence of surface objects. The proposed method can be used to track naval targets by unmanned aerial vehicles.