Primena inteligentnih sistema mašinske vizije autonomnog upravljanja železničkim vozilima

The railway is an important type of transport and has a significant economic impact on the industry and people's everyday life. Due to its capacities and complex infrastructure, it is necessary to work on its constant development and improvement. Railway automation requires the use of intelligent systems as a necessary part of an autonomous railway vehicle. As from the point of view of safe traffic, the existence of the object on the rail track and / or in its vicinity represents a potential obstacle to the railway traffic, and visibility has a very important role in correct and timely detection of the object on the railway infrastructure, a key element of autonomous railway vehicle is an obstacle detection system on the part of the railway infrastructure, in conditions of reduced visibility. The subject of scientific research of this doctoral dissertation is the application of intelligent machine vision systems in autonomous train operation. For the purpose of detecting obstacles on the part of the railway infrastructure in conditions of reduced visibility, a thermal imaging camera and a night vision system are integrated into the system, coupled with a developed advanced algorithm for image processing with artificial intelligence tools. In addition, the distance from the machine vision system to the detected object was estimated. The operation of the system was tested in a series of field experiments, at different locations, in different visibility conditions and weather conditions, through realistic scenarios