摘要
对快速路车辆异常行为的检测有助于防止或及时处理交通事故,缓解交通拥堵,保障出行的安全和效率。采用多高斯背景模型提取前景运动车辆及其中心点并利用Kalman滤波算法跟踪运动车辆。在此基础上,得到各个车道上车辆的行驶时空图,通过车辆时空图对车辆行为进行轨迹分析,根据时间序列上车辆位置的变化检测车辆逆行,通过车间距和车辆位置状态信息检测车辆碰撞。实验表明,该方法能较好地识别出车辆异常行为。
The detection of vehicle＇s abnormal behavior on freeway helps prevent or clear traffic accidents timely,relieve traffic jam and ensure traffic safety and efficiency.This paper applies a mixture Gaussian background model to extracting moving vehicles and their centers and uses the Kalman filter algorithm to track them.On this basis,the space-time diagram for vehicles running within each lane is obtained.Then,the trajectories of the vehicles are analyzed to detect wrong-way movements and vehicle collisions via vehicle spacing and location information.Study results clearly show that the proposed method can effectively identify the abnormal vehicle behaviors.
出处
《交通信息与安全》
2012年第4期89-92,98,共5页
Journal of Transport Information and Safety
关键词
车辆时空图
KALMAN滤波
异常行为检测
vehicle space-time diagram
Kalman filter
detection of abnormal behavior
