基于运动矢量的行车障碍自动检测算法研究

为了实现对车辆前方行人及车辆等威胁车辆正常安全行驶的危险障碍的检测,达到保证行车安全的目的,文章根据MPEG视频码流中运动矢量的特点,提出了一种新的行车障碍自动检测算法。该算法利用成熟的MPEG编解码技术对车辆前端摄像机所采集的视频图像序列进行编解码,输出MPEG码流格式的视频图像序列,通过分析解码的MPEG视频码流中运动矢量的规律,来制定判别行车前方障碍物的规则,并最终确定危险区域的范围。算法仿真结果表明,该算法具有较好的实践效果和一定的实用价值。

Research of Automatic Detection Algorithm of Traffic Barrier Based on Motion Vectors

In the purpose of realizing the detection of the threat to vehicle including pedestrians and vehicles and ensuring safe driving.According to the features of motion vectors in MPEG video code stream,this paper propose a new traffic barrier detection algorithm.The algorithm uses the mature MPEG codec technology to code and decode the video image sequence which is collected by the camera in the front of the vehicle,then output the video image sequence of MPEG stream format.Through the analysis of the law of motion vectors in the MPEG video decoded stream,to determine the rules of traffic barriers discriminant driving ahead,and finally determine the range of the dangerous area.The result of simulation shows that the algorithm has good practice effect and certain practice value.