【摘要】 为了实时监控公路隧道内车流量和交通事件,根据隧道特征、交通流变化情况、检测数据的相关性和常发事故的特点,在对目前视频检测系统分析的基础上,针对隧道内车流量和交通事件的检测,提出了一种基于多目标跟踪的隧道交通流视频检测算法。该算法根据当前图像帧的目标列表与前帧目标区域的目标列表进行相似性比较,计算其相似度以求出最小距离,这样减少了跟踪的盲目性,提高了算法的计算速度和精度。通过对隧道交通数据挖掘分析,提取隧道交通事件发生时各属性数据的关联性,试验测试表明对隧道车流量视频检测的总正确率高于97%及交通事件检测准确率高于95%。该算法能够在一定程度上对交通事故进行预警,对于减少和避免交通事故发生具有重要意义。更多还原

【Abstract】 In order to real-time monitor traffic flow and incidents in highway tunnel,based on tunnel characteristics,traffic flow situation,relevance of testing data and characteristics of common accidents as well as the analysis of current video detection systems,for detection of traffic flow and incidents in tunnel,a tunnel traffic flow video detection algorithm based on multi-target tracking was proposed.The algorithm compares the similarity of the target list of the current image frame and the target list of the former image frame,and then calculates the similarity and the minimum distance.Thus the tracking of blindness could be reduced and the computational speed and accuracy could be improved.Through data mining analysis of the tunnel traffic data,the correlation properties of the tunnel traffic incident attribute data were extracted.The test shows that the total video detection accuracy of tunnel traffic flow is more than 97% and the detection accuracy of traffic incidents is more than 95%.The algorithm can be used for early warning of traffic accident to a certain extent and it is of great significance for reducing and preventing traffic accidents.更多还原