摘要
以现有高架道路视频监控系统拍摄的视频作为数据来源,提出一种夜间拥堵判别方法。在中间车道设置虚拟线圈,通过监测车辆灯光对检测区域图像特征的影响识别车辆目标。若车辆增多光场复杂,以等时间夜间视频车辆目标检测作为补充方案,保障算法的实时性和准确度。若判断高架道路达到临界密度,短时禁行部分或全部高架道路上行匝道,同时将公共服务类车辆调离高架桥,规避交通拥堵。交通拥堵成因复杂,排堵过程难度大、耗时长,预测拥堵调度能有效缓解上述矛盾,同时提高视频监控系统的智能性,节约社会资源。
A method for judging congestion is proposed.The data source is night video,which is shot by the existing viaduct video surveillance system.The virtual coil is set up in the middle lane to identify the vehicle target,and the influence of vehicle light on image features is monitored.If the light field of the vehicle is complex,the vehicle target detection at night is used as the supplementary scheme to ensure the real-time and accuracy of the algorithm.If it is judged that the viaduct reaches the critical density,some or all of the viaduct ramp will be banned for a short time,and the public service vehicles will be removed from the viaduct to avoid traffic congestion.Because of the complexity of traffic congestion and the difficulty of removing congestion,the prediction of congestion scheduling can effectively alleviate the above contradictions,improve the intelligence of video surveillance system and save social resources.
作者
刘阳
陈荣保
LIU Yang;CHEN Rongbao(School of Electrical and Automation Engineering,Hefei University of Technology,Hefei 230009,China)
出处
《传感器与微系统》
CSCD
2020年第12期57-60,共4页
Transducer and Microsystem Technologies
关键词
夜间视频检测
高架道路
车辆目标识别
预先排堵调度
night video detection
viaduct
vehicle target identification
advance congestion scheduling