Review of modern technologies of vessel monitoring in the water areas of the inland waterways using video surveillance tools

Abstract

At present, there is a problem on inland waterways located within the boundaries of large settlements to ensure monitoring of vessel traffic, including small vessels. This situation is due to limitations on the use of traditional means of monitoring of inland waterway vessels, caused primarily by restrictions on the use of radar systems and radiotechnical means operating in the ultrashortwave frequency range within the city limits. At the same time in large cities of the Russian Federation implemented and successfully used video surveillance systems. The system operating in the city of Moscow to monitor the water area within the city as part of the city traffic control system is considered. Most of the sections of inland waterways located within the limits of St. Petersburg fall within the range of the cameras. The article considers possible approaches to use the city video surveillance system for monitoring the water area within the city limits of St. Petersburg. The paper describes technologies based on artificial neural networks potentially suitable for identification of ships and determination of their exact location at a given moment of time. Advantages and disadvantages of the considered methods are analyzed, as well as a variant of solving the inverse problem of bearing of ships (from the shore) using video surveillance systems consisting of two cameras is proposed. The algorithm of the system operation for identification and determination of vessel movement parameters is proposed. Domestic and foreign experience of solving the problems of vessel identification using video surveillance systems, as well as determining the parameters of vessel movement is analyzed. Several neural networks (object detection, text recognition) are proposed as a solution as a basis for further study of the described problem.