COMPARISON OF ROAD TRAFFIC MEASUREMENT METHODS BASED ON VIDEO MATERIAL ANALYSIS ? AN OPTICAL FLOW METHOD AND GAUSSIAN MIXTURE MODEL

Resumen

The article presents two selected road traffic measurement methods based on video material analysis. A comparative analysis in terms of their accuracy, usability and comfort of use was made. Parameters detection accuracy in the video material depend on a range of variables. Fundamental is the measurement method selection. It has been proposed to measure the number of vehicles ? using two methods. The first one ? a method of optical flow ? allows to detect the parameters of passing vehicles covered by the sequences of film. The choice of above methods was determined by possesion of the MatLab as well as key extension ? Imaging and Vision Processing Toolbox. In the context of studies the video material was analyzed using mentioned methods. The results were also compared with the actual state calculated ex post facto from the video material, without use of computational algorithms.