视频序列中目标的检测与跟踪

摘要

视频序列中,为降低背景变化对目标的检测与跟踪产生的影响,提出了一种新目标检测与跟踪算法.首先,以像素为基础建立时间序列背景模型,提高了目标检测的精度.然后,采用KD-tree算法进行目标跟踪,获得了运动目标的完整路径.实验结果表明,该方法在视频序列目标的检测与分析中,适应环境变化的能力强,达到实时处理的效果.%In video sequence, a new algorithm of target detection and tracking is proposed for reducing the impact of background change on the target detection and tracking. Firstly, time series background model, which is based on pixel, is built to improve the target detection precision. Secondly, the complete path of the moving target is achieved by using KD-tree algorithm. The experimental results show that this model have strong adaptability to environmental change and real-time in the target detection and analysis of video sequence.