摘要
基于视频图像的车辆运动跟踪是当前计算机视觉研究的热点,具有广泛的应用领域。本文研究了一种基于视频图像的运动车辆跟踪算法,通过当前帧目标边缘与实时更新模板的最优匹配来确定目标的位移量。首先启动了卡尔曼滤波,预测目标匹配搜索区域,然后再在搜索区域中利用边缘匹配精确定位目标,减小了匹配的计算量。实验结果表明,该方法是有效的且实时性好。
Tracking the motion of vehicles based on video sequence images is one of the hot spots in computer vision-research, which has broad application. This paper presents a tracking algorithm based on video sequence images. The edge matching determines the optimal displacement vector the target edge template and the current edge, and the template is u between two successive frames by matching of pdated in real time. This method first use Kalman filter to predicate the search areas of the matching objects, then match the edge in correct areas to track the vehicles. Therefore the computation is greatly reduced. And the experiments result shows that this method can track the vehicles effectively and get a better time-consuming.
出处
《电子测量与仪器学报》
CSCD
2009年第3期45-48,共4页
Journal of Electronic Measurement and Instrumentation
基金
国家自然科学基金(编号:20299030)资助项目
扬州大学自然科学基金(编号:KK0313090)资助项目
关键词
视频图像
卡尔曼滤波
边缘匹配
车辆跟踪
video sequence images
Kalman filter
edge matching
vehicle tracking
