论文部分内容阅读
车辆跟踪中普遍存在车辆遮挡将直接影响事件检测精度的情况。为了解决这一问题,在介绍车辆跟踪算法基本原理的基础上,提出了一种基于视频与检测器数据组成的语义层次交通事件检测算法。该算法先应用时空马尔可夫随机场模型进行车辆跟踪,得到交通流基本参数,然后结合安装在道路下游的检测器获得的交通流数据一起,采用语义层次算法对交通事件进行检测。为了验证算法的准确性,最后对该算法与不使用检测器数据算法进行比较,发现使用检测器数据算法的检测率要高。通过研究得出:基于视频与检测器数据组成的语义层次算法在交通量比较拥挤且车辆出现相互遮挡的情况下,能准确判断交通事件发生,对缓解交通拥挤和减少交通事故有重要的意义。
Vehicle tracking is widespread vehicle occlusion will directly affect the event detection accuracy of the situation. In order to solve this problem, a semantic-based traffic incident detection algorithm based on video and detector data is proposed based on the basic principle of vehicle tracking algorithm. The algorithm firstly uses the space-time Markov random field model to track the vehicle and obtains the basic parameters of the traffic flow. Then, the traffic level is detected by using the semantic hierarchy algorithm combined with the traffic flow data obtained by the detectors installed in the downstream of the road. In order to verify the accuracy of the algorithm, we finally compare the algorithm with the algorithm that does not use the detector data and find that the detection rate using the detector data algorithm is higher. Through the research, it is concluded that the semantic hierarchy algorithm based on the video and the detector data can accurately determine the occurrence of traffic accident when the traffic volume is relatively crowded and the vehicles occlude each other, which is of great significance to alleviating traffic congestion and reducing traffic accidents.