中文摘要：

经典Mean Shift目标跟踪算法采用单一颜色特征建立目标模型,在目标颜色与背景颜色相近或遮挡的情况下,目标跟踪鲁棒性差,为此,提出另一种Mean Shift目标跟踪算法。采用Haar型局部二值模式（Haar local binary pattern,HLBP）算子提取HLBP纹理特征,利用HLBP纹理特征具有较强辨识度、对光照变化不敏感等特点,代替原始视频图像序列,建立HLBP纹理特征的空间概率密度分布模型来表征目标特征;在此基础上,在Mean Shift框架下获取目标位置估计值,实现目标的跟踪。对比实验结果表明,该算法具有较高的目标跟踪精度和鲁棒性。

英文摘要：

To improve the poor performance of the Mean Shift tracking algorithm based on single color feature that adopted to establish the target model when colors of target and background are similar and shaded,Mean Shift target tracking algorithm based on texture feature was presented.Texture feature was extracted using Haar local binary pattern operator（HLBP）,considering the advantages including strong identification and insensitivity to illumination change of HLBP feature,HLBP texture feature figure was used to replace the original video image in the calculation process.HLBP texture feature probability density distribution model was established to characterize target feature.The target location was estimated in the framework of Mean Shift.Experimental results show that the proposed algorithm can improve the accuracy of target tracking effectively.