Moving Objects Detection Technology Based On Real-time Traffic Flow

In recent years, the transportation network has been vigorously built all over the world, which has brought convenience in traveling and other aspects of life. At the same time, it comes along with frequent traffic accidents, which have brought heavy suffering and huge economic losses. It’s of limited effect to solve the traffic problems only by traditionally widening the roads and improving the transport facilities. It has recently become a research hotspot to strengthen the real-time monitoring and management of the traffic flow through intelligent traffic management, which have significant effect on solving the problem of frequent traffic accidents, and reduce personnel and dispatch the traffic flexibly.The parameters detection of traffic flow in the intelligent transportation systems is focused in the paper, which is studied in the fields of artificial intelligence, pattern recognition, image segmentation, edge detection and mathematical morphology, etc. The main works are as follows:1, Methods study on background model description. Selection of parameters, expressions and noise effect of Local binary pattern (LBP) being analyzed, binary LBP operator is presented and improved by an anti-noise factor being introduced.2, Study on integration background modeling method. Modeling research on texture information and color information of background, a background modeling method integrated color information and texture information is presented. Image segmentation technology and morphological image processing technology are introduced to improve the detection results of moving objects as well.3, Study on movement tracking. Sector block search and Camshift algorithm are studied, and a tracking algorithm matched forecast of integration region with Camshift region, by which the objects are tracked through region forecast, calculation of color probability distribution of the image, and judgment of if the regions being matched.4, Study on detection of traffic parameters. A framework of detection of traffic flow parameters is presented. And an average speed detection model and an instantaneous speed in ideal state detection model are presented through the research on traffic flow and speed in the detection of traffic flow parameters. To sum up, intensive research is taken centre on detection of traffic parameters, and relevant solutions or algorithms are presented, whose effectiveness is verified by plenty of experiments.