摘要
在基于视频的交通监控系统中,车辆的快速、有效提取是车辆检测中的一个重要环节。在彩色图像中,当车辆出现在道路上时,因为车辆与道路相比颜色上有较大的差异,所以在车辆的边缘像素点上会发生颜色的突变。本文以此为依据,结合车辆的几何信息,提出了一种新的车辆提取算法。此方法通过对彩色图像进行逐行扫描的方式,利用滑动差分滤波器,确定每行的车辆边缘像素点,进而将车辆从图像中提取出来。实验证明,该方法能简单有效地提取车辆。
For a video-based traffic monitoring system, extracting vehicle in a fast and effective way is an important step for vehicle detection. When the vehicle appears on the road, the pixels of the vehicle＇s edge will have a sudden change in color because of the difference in color between the vehicle and the road in the color image. This paper proposes a novel vehicle extracting method based on the fact mentioned above and the information of the vehicle＇s structure. The changing degree of color feature is also used. This method can locate the position of the vehicle＇ s edge pixels by scanning each line of the image and using moving-difference filter, and extracting the vehicle furthermore. The experimental result shows that this method can extract the vehicle in a simple and effective way.
出处
《电子测量技术》
2008年第11期16-20,共5页
Electronic Measurement Technology
