Traffic Flow Detection Technology Based On Aerial Video

As the economy continues to develop, an increasing number of vehicles, thisworsening traffic environment. Establish an efficient comprehensive intelligenttransportation systems has become a top priority for traffic management. Traffic flowinformation detection system is intelligent transportation systems to improve the lot ofbasic traffic flow information is an important part of the Intelligent TransportationSystem..With the development of machine vision and image processing technology, thevideo-based traffic detection technology has increasingly become mainstream trafficdetection technology for the industry concerned. This paper analyzes the status quo at thisstage study, a detailed analysis describes the video-based traffic detection technology inthree key components: vehicle detection, statistical tracking of vehicles and vehicleoperation data. This paper mainly three issues as the main entry point, proposed andestablished a set of processes and algorithms based on aerial video traffic detectionsystems.In this paper, the vehicle detection background subtraction method, proposed a newadaptive weights in the background modeling algorithm, and by background subtraction,binary, morphological processing and a series of processes to achieve the detection ofvehicles.In the background modeling, by adjusting the value of several factors that canadapt to the specific requirements of different situations, strong adaptability.In tracing the vehicle tracking method using a feature-based, mainly in the center ofmass of the vehicle as well as the length and width of a rectangle whose externalcharacteristics, establish similarity function to achieve the integration of multi-featurematching vehicle. With feature matching basis, this article lists the region of interest ROIapproach by setting track to achieve the stability of the vehicle tracking.In terms of traffic flow information, on the basis of the vehicle on the track to achievea variety of determining the direction of travel of the vehicle, and the vehicle travelingspeed judgment, but given the pressure for road traffic record. Experiments show that the proposed method is proved, the algorithm simple, strongaccuracy, real-time speed is good, practical, intelligent transportation system to meet thedemand.