移动视频目标跟踪中的智能过滤算法研究

Research on intelligent filtering algorithm in mobile video target tracking

随着视频监控系统在现代生活中的应用越来越广泛,对移动视频目标跟踪算法的研究也越来越重要。由于视频目标追踪非常复杂和困难,如何提高移动视频目标跟踪的鲁棒性、准确性和实时性成为当前发展的主要方向。本文通过视频目标跟踪理论分析,并通过改进的目标跟踪过滤算法,解决了原来Lucas Kanade光流跟踪算法的目标易丢失、鲁棒性不高、计算量大等缺点。实验结果证明,智能过滤算法能够准确预测目标质心位置,实现可靠的跟踪。

With the video monitoring system in the application of modern life more and more widely,mobile video target tracking algorithm research is becoming more and more important. Due to the complexity and difficulty of video target tracking,how to improve the robustness,veracity and instantaneity of mobile video target tracking become the main direction of the current development. In this paper,through theoretical analysis of video target tracking,and through improved target tracking filtering algorithm has solved the some shortcomings of the original Lucas Kanade optical flow,such as target easily lost,robustness of the algorithm is not high,large amount of calculationand so on; Through the experimental results show that intelligent filtering algorithm can accurately predict target centroid position,then achieve reliable tracking.