基于多线索融合的目标跟踪算法研究  

A Target Tracking Algorithm Research Based on Multi-Cue Fusion

作　　者：谢之宇[1] 蒋晓瑜[1] 汪熙[1] 裴闯[1] 

机构地区：[1]装甲兵工程学院控制工程系,北京100072

出　　处：《计算机技术与发展》2011年第3期125-127,131,共4页Computer Technology and Development

基　　金：军队"十一五"预研项目(404010204)

摘　　要：目标跟踪是计算机视觉领域一项核心技术,通用的目标跟踪算法以单线索或简单融合多个线索为主,当背影突然改变时,单一线索或多个线索的简单融合便难以跟踪成功。针对这个问题,在粒子滤波的框架下提出了自适应融合颜色线索和角点线索的跟踪方法,通过判断各个线索的可信程度,自适应给不同线索分配不同的权重,很好地解决了在复杂背景、互相遮挡情况下的跟踪问题。实验证明,采用的自适应多线索融合方法在实际应用中有更强的鲁棒性。Target tracking is the key technology on the field of computer vision. Methods that generally used now are mainly based on single cue or straightforwardly fusing several cues. So when the background suddenly changed, this methods are hard to track the target. In order to handle this problem, a tracking algorithm based on adaptive combination of color cue and comer cue was proposed to solve the tracking problem under complex situation properly by estimating the reliability of each cue and distribute different weight to different cues automatically. The experiment results prove that the self-adaptive multi-cue fusion algorithm is stable and robust in the application.

关 键 词：目标跟踪 颜色线索 角点线索 多线索融合 自适应融合 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]