Interval Speed Detection Combined with Vehicle Key Points Recognition 2019-01-0919
Speed monitoring is an important way to ensure traffic safety. The characterized dynamic speed monitoring can effectively analyze the automobile driving characteristics and the traffic flow, which is helpful to develop intelligent transportation system. The radar speedometer’ scope of the speed application is beyond 5km/h and the loop detector is difficult to accurately judge the speed when the distance between two vehicles is less than three meters. This paper proposed a vehicle speed detection method combining vehicle key points recognition. The method consists of two main phases, including the key points extraction and the speed fitting. Firstly, at the stage of key points, the target vehicle is extracted through image processing, and four key points of the target vehicle are obtained. These four key points are respectively used to represent the four scenes of day, night, rainy day and snow day. At the speed fitting stage, after the running track of the key points are fitted, the distance and time data will be obtained. And then the weight of the fitting data is distributed according to the scene information and the definition of the key points, so that the main weight is transferred to the key field scenic point. After weighting the fitting data, the running speed of the target vehicle is acquired. This method realizes the real time detection of vehicle speed, and has the characteristics of wide range of velocity measurement and multiple working conditions, which is of certain guiding significance for the research of intelligent transportation system.
 Shao fangfang. An Overview of Traffic Detector.Shandong Jiaotong University.2014