一种改进的均值移位红外目标跟踪算法

为了增强复杂背景条件下红外目标跟踪的稳健性，提出了一种改进的均值移位目标跟踪算法。该算法融合了均值移位的梯度匹配搜索策略的优势与基于特征分类跟踪算法强鲁棒性的优点，建立了灰度似然比加权的核直方图目标表征模型。模型中加入了目标与局部背景灰度特征的似然比作为原始核直方图的权值，应用该模型的均值移位算法能够进一步提高目标像素灰度的移位权重，有效抑制背景干扰，进而提高低对比度目标跟踪的稳健性。同时，基于跟踪复杂度估计提出了目标遮挡情况下的模型更新判别准则，提高了算法的自适应性能。实测红外目标跟踪实验表明了该算法简单、有效。

An improved mean shift based IR target tracking algorithm

In order to enhance the robustness of IR target tracking under complex background, an improved Mean Shift tracking algorithm is proposed. This algorithm combines the advantage of gradient matched searching strategy based on mean shift with the robustness of the tracking algorithm based on featutes classification，and likelihood ratio weighted kernel histogram target representing model is presented. In improved model the likelihood ratio of gray level feature of target and local background is regarded as a weighted value of original kernel histogram, and then the robustness of tracking the IR target with low contrast is enhanced by increasing the shift weighted value of target gray level in the form of mean shift tracking, and background interference is suppressed effectively. Adaptive performance of the algorithm is improved via proposing the discrimination criterion of model updating based on tracking complexity estimation under target occlusion. The validity of new algorithm is verified by the actual experiment.