摘要
为了提高智能视频监控系统中运动目标跟踪的准确率和有效性,利用数据融合和协同跟踪技术,实现了多摄像头协同跟踪系统中运动目标的有效跟踪。首先,利用单摄像头独自跟踪得到目标图像,然后利用平面单应性和极线几何约束结合实现多摄像头视图间的目标匹配和协同跟踪,最后得到精确的目标跟踪结果。通过在MATLAB上实验仿真,并与单摄像头目标跟踪进行对比,得出该算法具有较好的跟踪效果以及抗遮挡性能,可广泛应用于智能视频监控系统中。
To improve the accuracy and effectiveness of the moving objects tracking in intelligent video surveillance system, using data integration and collaborative tracking technology, effective tracking of moving targets in multi-camera traeking system is achieved. Firstly, the foreground image is got by every single camera targets tracking, and then use the planar hnmography and the epipolar constraint,targets correlation and collaborative tracking between muhi-camera views are achieved, and finally accurate tracking results are got. Through experiments and simulations in MATLAB, comparing with single camera tracking results, the algorithm has better tracking performance and anti-blocking properties, and can be widely used in intelligent video surveillance system.
出处
《电视技术》
北大核心
2015年第18期86-89,98,共5页
Video Engineering
基金
山西省自然科学基金项目(2013011017-3)
关键词
智能视频监控
数据融合
协同跟踪
平面单应性
几何约束
intelligent video surveillance
data integration
collaborative tracking
planar homography
epipolar constraint
