一种单幅运动模糊图像的车速测量算法

A vehicle speed measurement algorithm based on single motion-blurred image

针对道路中行驶车辆速度测量的问题,利用运动车辆所产生的单幅运动模糊图像可获取车辆在曝光时间内的运动信息,从中提取其运动参数,进而实现行驶车辆的速度测量。在分析运动模糊成像机理和车速测量模型的基础上,构建道路交通场景实验环境,用CCD相机以多种快门速度对车速在25~130 km/h之间的行驶车辆进行多次抓拍;通过定位单幅模糊图像中车辆轮毂的边界,并计算轮毂运动方向和非运动方向上的直径差来估算车辆的运动参数,完成车速测量实验。实验数据表明:不同曝光时间和不同车速情况下,车速测量的误差绝对值都在8%以内。结果表明该算法对成像环境要求不高,且适用于不同车型。

With regard to speed measurement of running vehicles on road,a single motion-blurred image of moving vehicle is used to obtain the moving information of the vehicle the during exposure time so as to extract its motion parameters and finally work out the speed vehicle of.After analyzing the motion blur imaging mechanism and vehicle speed model measurement,a test environment of traffic scene on the road is set up.And then,a CCD camera used is to capture the images of moving vehicles running at a speed of 2 5-1 3 0 km/h at different several shutter speeds.After that,by positioning the border of the vehicle’s wheel hub in blurred the image and calculating the diameter difference of the wheel hub in the moving and direction the non-moving direction,the motion parameters of the vehicle are calculated to finish vehicle the vehicle the speed measurement.Experimental data shows that the absolute value of error of speed that speed'showing the algorithm doesn’t have high requirements on imaging environment and it is suitable different for kinds of vehicles.