Sažetak
The Faculty of Transport and Traffic Sciences Computer Vision Group conducts research regarding theory and application of image processing for transport and traffic related problem solving. Group members are from different technical areas (traffic science, transport technology, computer science, control, robotics) ensuring high interdisciplinarity benefits for problem solving. Main research problems include origin-destination matrices extraction from intersection video footage, video geo-referencing for road evaluation, traffic sign detection including georeferencing, automatic traffic incident detection, and hyperspectral imaging applications in protection of transport environment. Research results obtained individual and in collaboration with other research groups also include image segmentation and feature extraction, object recognition and tracking, and video geo- referencing. Research equipment includes a road evaluation vehicle (equipped with HD cameras, GPS and retroreflectometer), accessibility to multi- sensory airborne imagine system (equipped with thermal, multispectral and hyperspectral sensors), and software package for license plate recognition. Group has also access to an test area with road, parking and airfield infrastructure enabling testing of developed applications in real world conditions. Current and future work is related to traffic parameters measurement and estimation from video-footage obtained using unmanned aerial vehicles, automated computer vision based road safety analysis, and computer vision based parking lot management evaluation in an outdoor test site.