基于视频虚拟检测带的交通流量检测方法

交通数据的正确掌握对后续的交通控制起着至关重要的作用,采用基于视频虚拟检测带的交通流量检测方法,对检测带范围内像素点RGB的强度值进行统计,将无车辆时的统计值作为标准特征模板并执行模板更新策略,通过计算每一帧实时视频图像的特征值与标准模板特征值的距离进行车辆检测.为了提高检测精度,提出了一种基于阴影特征的阴影检测算法.在早晨和中午的不同光照条件下进行试验,在阴影明显的情况下,该方法的检测准确率在91％以上,在阴影不明显的情况下,检测准确率达到95％,可满足智能交通的需要.

Traffic Volume Detection Methods Based on Video Virtual Strip

The acquirement of precise traffic information plays an crucial role in subsequent transportation control.A traffic volume detection method based on video virtual strip is put forward.RGB values in the range of virtual detection strip are calculated,a distribution characteristic of no vehicle detected is used as the standard template and to execute template update strategy.The difference between real-time characteristics of every frame image and the standard characteristics is calculated for vehicles detection.A shadow detection algorithm based on shadow feature is proposed to improve the detection accuracy.Experiments are carried out under different conditions in the morning and noon,the recognition rate is higher than 91% with obvious shadows,while the recognition rate is even higher than 95% without shadows.The detection results can satisfy the needs of intelligent transport system.