论文部分内容阅读
为了对交叉口车辆的位置进行准确定位,提出了一种分布式视频网络架构下车辆精确定位方法。在分布式视频网络中每处摄像机架设位置均设有2类摄像机:近景摄像机和远景摄像机。首先在近景摄像机拍摄范围内,对感兴趣区域内车辆进行身份识别,根据车牌照平面与道路平面垂直的约束条件,建立车牌照模型来对车辆精确定位;接着在远景摄像机拍摄范围内,采用融合局部二值模式(LBP)纹理特征的金字塔稀疏光流法实时跟踪车辆上局部特征点,根据特征点运动趋势相似性获得稳态特征点,来对车辆位置估计;最后根据不同摄像机检测结果,采用加权一致性信息融合算法来提高车辆定位精度。实验结果表明:该方法能对交叉口车辆位置进行精确定位。
In order to accurately locate the position of the vehicle at the intersection, a precise positioning method for vehicles under distributed video network architecture is proposed. There are two types of cameras at each camera erection site in the distributed video network: close-range cameras and long-range cameras. Firstly, the vehicles in the region of interest are identified within the shooting range of the close-range camera, and the license plate model is set up to accurately locate the vehicle according to the constraint conditions that the license plate plane and the road plane are vertical. Then, within the range of the long-range camera shooting, The local binary pattern (LBP) texture feature pyramid sparse optical flow method is used to track the local feature points in real time and obtain the steady state feature points according to the similarities of the feature point motion trend to estimate the vehicle position. Finally, according to different camera detection results, Weighted consistency information fusion algorithm to improve vehicle positioning accuracy. Experimental results show that this method can accurately locate the vehicle position at the intersection.