摘要
针对复杂背景下,尤其是当光照条件发生变化以及目标发生遮挡时容易导致跟踪失败的问题,提出了一种基于自适应多特征融合的粒子滤波目标跟踪算法;该算法将RGB颜色直方图和LBP纹理直方图融合起来建立目标参考模型,并且引入Sigmoid函数动态调整两类子特征粒子的权重;仿真结果表明,该算法能在复杂背景下自适应调整两种子特征权重,以克服其中一种特征失效导致的跟踪失败,而且有效地避免了使用单一特征建模的缺点,能够实现更加准确的跟踪。
An adaptive tracking algorithm combining multi--cue based on particle filter is proposed to overcome the disadvantages of complex background, especially when the light changes or the target gets blocked, can easily lead to tracking failure. A color histogram and a texture histogram were combined to build the object＇ s reference model. Sigmoid kernel function is used to regulate the feature weight dy- namically. Simulation results demonstrate that, the tracker based on multi--cue particle works can adjust the two kinds of feature weights a daptively, effectively overcoming the disadvantages posed by modeling with a single feature and improving the accuracy of object tracking.
出处
《计算机测量与控制》
北大核心
2014年第4期1182-1184,1188,共4页
Computer Measurement &Control
基金
国家自然基金项目(61172185)
关键词
目标跟踪
粒子滤波
颜色特征
object tracking
particle filtering
color feature
