基于扩展空间直方图的红外目标跟踪方法

Infrared Object Tracking Algorithm Based on Extended Spatial Histogram

针对传统均值漂移算法中仅仅利用目标的颜色信息而导致目标模型分辨能力不高的问题,提出了一种基于扩展空间直方图的红外目标均值漂移跟踪方法.首先对空间直方图进行扩展,构建了一种结合目标颜色分布和空间约束关系的联合空间颜色模型,有效提高了目标模型的分辨能力.通过给定目标空间位置和颜色联合概率密度函数,定义目标区域与候选区域概率密度的相似性度量,进而实现了红外目标的准确定位.实验结果表明该算法简单有效,能准确跟踪前视红外目标.

To improve the low discriminative ability of object model with only color information in traditional mean shift algorithm, improved mean shift tracking algorithm using spatio and gray features is proposed. Firstly, through extending the spatio histogram, the joint spatio-gray object model is constructed with both gray distribution and spatio relationship. Meanwile, the joint probability density function is defined to build the similarity measure between the object model and candidate model. Finally, the accurate localization of infrared object is realized by the extended mean shift algorithm. Experiment results verify the effectives and robustness of the proposed algorithm which can improve the tracking performance efficiently.