基于视频投影法的交通拥挤实时检测算法  

Video-based projecting algorithm for real-time traffic congestion detection

作　　者：张新[1] 常云涛[1] 郭佳宁[2] 谢平红[1] 

机构地区：[1]同济大学交通运输工程学院,上海201804 [2]重庆交通大学交通运输学院,重庆400074

出　　处：《合肥工业大学学报（自然科学版）》2013年第3期319-323,共5页Journal of Hefei University of Technology：Natural Science

基　　金：国家高技术研究发展计划(863计划)资助项目(2009AA11Z220)

摘　　要：针对日益严重的城市道路交通拥挤问题,文章提出了基于视频检测技术实时判断道路交通拥挤的算法。首先通过视频投影法将二维车道转化为一维线段,再通过扫描线段方法获取车道占有率,然后用指数平滑公式对提取出的车道占有率进行去噪处理,接着以道路占有率、占有率方差为交通特征参数,建立交通拥挤识别模型,最后使用模型对3段实例视频进行处理,用模型识别出的结果与人工判断的结果进行对比分析,得出使用模型进行拥挤识别的正确率在92%以上,证明了模型的有效性。Aiming at the growing problem of urban traffic congestion,a video-based projecting algorithm to detect the real-time traffic congestion on the urban road is proposed.Firstly by video projecting method,two-dimensional lane is transformed into one-dimensional line segment,the scanning line method is used to obtain lane occupation rate,and the exponential smoothing formula is used to remove the noise in the source data.Then considering the lane occupation rate and its variance as the traffic characteristic parameters,the traffic congestion detection model is established and used to process three sections of traffic video.The results detected by the model are compared with those detected by artificial judgment,and it is showed that the correct detecting rate is above 92% by the proposed algorithm,thus demonstrating the validity of the algorithm.

关 键 词：智能交通 视频投影法 交通拥挤识别 城市道路 实时检测算法 

分 类 号：U441.1[建筑科学—桥梁与隧道工程]