Lobus R. Traffic system sensor video processing methods

The scientific and technical task decision of algorithmic and hardware providing traffic monitoring in difficult Ukraine traffic parameters is devoted in this thesis. The more perfect videoprocessing methods and algorithms of the automatic videosupervision systems are offered, after the amount by calculations and functioning efficiency. The curvilinear co-ordinates system is used in the objects tracking process for calculations amount diminishing. A method, reverse to Brezenkhem method, allows to execute the scan-out of image in this system of co-ordinates. Developed algorithms of determining the amount, speeds, and types of transport vehicles complement the developed videostream scan-out method, that proves it functional possibilities plenitude. Hardware research results is a traffic sensor which gives possibility to realize the automatic or automated vehicles parameters collection systems with traffic control in big cities. Traffic sensor functioning efficiency experimental researches are conducted. Hardware decisions, functionings on the developed methods and algorithms base, are inculcated on the state and private enterprises of Ukraine.