面向车辆视频测速的摄像机倾斜角自适应标定方法

Adaptive calibration of camera tilt angle used for video speed measurement

车辆视频测速中常常需要标定摄像机安装高度、视场角、倾斜角等参数,而其中倾斜角的确定与更新一般采用人工方式,限制了视频测速系统的应用。针对此问题,提出了一种自适应的摄像机倾斜角标定方法,利用连续观测图像序列建立超定线性方程组,并采用最小二乘法进行求解,最后经过闭环反馈调节,可以实时得出摄像机倾斜角。实验结果表明,该方法测量较为准确,避免了传统人工测量和维护的不便,具有较好的实用性。

Vehicle speed measurement through video often need a calibration of mounting height,field angle,tilt angle.But generally uses artificial means to determine and update the tilt angle,and it limits the application of video speed measurement system.To address this issue,this paper proposed an adaptive calibration of camera tilt angle,which established overdetermined linear equations through consecutive observation image sequence,and used the least square method to solve,then finally through the evaluation and adjustment,it obtained real-time camera tilt angle.Experimental result shows that the measurement is more accurate,avoiding the inconvenience of manual measurement and maintenance,and it is good practical.