无人机广域视频的机动车交通参数计算及分析  

Calculation and Analysis of the Traffic Parameters of Vehicles Based on the Wide-area Drone Video

作　　者：周雨阳[1] 龚艺[1] 姚琳[1] 李小勇[2] 

机构地区：[1]北京工业大学北京市交通工程重点实验室,北京100124 [2]北京工业大学现代教育技术中心,北京100124

出　　处：《交通运输系统工程与信息》2015年第6期67-73,共7页Journal of Transportation Systems Engineering and Information Technology

基　　金：国家自然科学基金青年科学基金(51208014);北京市教委科技计划面上项目(KM201310005026);北京工业大学基础研究基金项目(X4106001201401)

摘　　要：基于视频的交通参数提取在智能交通系统中具有重要意义.针对无人机视频,提出了一种基于大范围无人机视频的机动车交通参数提取方法.首先通过建立无人机广域镜头图像的校正算法去除交叉口视频图像的广角畸变,然后将像素坐标转化为真实坐标.其次,利用帧间差分法计算行驶车辆的速度、加速度、车头间距等车辆运行参数,将计算结果与真实车辆及软件提取运行参数进行对比,结果证明该方法具有较高的准确性和可行性.最后,实例验证该方法能较准确地获得交叉口车辆运行参数,并将其应用于车辆通过交叉口全过程的速度变化规律分析.The video based traffic parameters extraction has great significance in intelligent transportation systems. For the drone video, a new method of the traffic parameter extraction is proposed in this paper. The correction algorithm of the wide- angle lens of the drone is established for eliminating the video images distortion of the intersection. The relation between the image coordinate and the true coordinate is built. Then the frame difference algorithm is used to obtain the traffic parameters of the vehicles, such as the speed, the acceleration and the distance headway. The high accuracy and feasibility of the algorithm is proved by the comparison of the real parameter, the parameter from this paper and that from the software. Finally, an example is given to verify the practicality of this method as well as to analyze the velocity regulation during the vehicles passing the intersection.

关 键 词：智能交通 无人机视频 校正算法 交通参数 坐标转化 帧间差分法 

分 类 号：U491.1[交通运输工程—交通运输规划与管理]