北京交通大学学报2011，Vol.35Issue(4)：138-141,147,5.
交通事件视频图像自动识别系统的研究
Study of traffic incident automatic recognition system based on video image
摘要
Abstract
Presently it is extremely urgent for highway's visual monitors to make full use of visual information on expressway, detect traffic incidents intelligently 24-hours, and raise first-aid treatment efficiency to the department of traffic management. Based on the camera set on the expressway,video image processing was used to detect the traffic incident. Firstly, in order to extract boundary of foreground objects and inhibit noise, the image pre-processing methods of median filter and image scaling were employed. Binarization segmentation of foreground objects was processed based on the threshold value algorithm. And then, a connected region labeling algorithm of binarization scene image was proposed to extract feature of foreground objects and detect the traffic incidents. Experimental results show that the algorithm can filter the waves and decline the noisy effectively. The system can recognize traffic incidents accurately and rapidly.关键词
交通工程/交通事件检测/视频图像/阈值法/连通区域Key words
traffic engineering/ traffic incident detection/ video image/ threshold value algorithm/connected region分类
交通工程引用本文复制引用
赵建东,张昊,王自上,李兆轩..交通事件视频图像自动识别系统的研究[J].北京交通大学学报,2011,35(4):138-141,147,5.基金项目
国家科技支撑计划项目资助(2011BAG071B05) （2011BAG071B05）
中央高校基本科研业务费专项资金资助(2011JBM094) （2011JBM094）
