基于视频的公路交通事件检测算法研究

Research on Video-based Highway Traffic Incident Detection Algorithm

本文针对当前公路交通事件检测算法,在识别交通事故时,对移动车辆图像纹理特征提取的稳定性较低的问题,设计了一种新的基于视频的公路交通事件智能图像检测算法,首先对公路摄像头采集的动态车辆视频图像进行预处理,利用环形线圈传感器与视频识别系统进行车辆图像数据采集,对数据视频图像进行数据处理以及灰度幅值特征、直方图统计特征的提取,通过稳态基本特征差异大小的判断,实现基于视频的公路交通事件检测算法,并通过实验验证了该算法的有效性。

Regarding the problem that the current highway traffic incident detection algorithm shows a low stability when extracting image texture features of a moving vehicle for identifying a traffic incident,a new video-based intelligent image detection algorithm for highway traffic incidents is designed. According to the detection algorithm,the dynamic vehicle video image collected by highway cameras is preprocessed at first. The image data of the vehicle is captured by the loop coil sensor and the video recognition system. The data processing is performed on the data video image,and the gray amplitude feature and the histogram statistical feature are extracted. The video-based highway traffic incident detection algorithm is realized by the judgment on the difference of basic features of steady state. The effectiveness of the algorithm is verified by experiments.