基于QP_TR信任域的序列图像目标跟踪

Tracing of Objects in Image Sequences Using QP_TR Trust Region Algorithm

本文将信任域算法和尺度空间理论相结合,提出了一种能够精确描述目标尺寸连续变化的新的序列图像目标跟踪算法;将信任域算法与灰度模板相结合,提出了一种新的实时目标跟踪算法。在第一种算法中,首先将序列图像按照颜色直方图转换成目标概率分布图,目标区域在概率分布图中呈现为灰度块。然后通过检测该图在尺度空间中微分滤波器输出的极值,来决定这些灰度块的尺度。最后我们使用QP_TR信任域算法在尺度空间里和图像平面内快速搜索概率分布图的多尺度规范化Laplacian滤波函数极值,实现了目标定位并同时决定了其尺度,从而完成了跟踪任务。在第二种算法中,首先记录目标初始模板,在随后每一帧中应用QP—TR信任域算法搜索与该模板最相似的区域,实现目标定位。和现有算法的比较以及在大量真实序列图像上的实验表明,两种算法分别在目标大小描述,跟踪精度上以及运算速度上有了显著提高。

A new tracking framework based on the QP-TR trust region algorithm is proposed, in which two independent algorithms appropriate for different situations are demonstrated. In the first algorithm the constant changes of the target＇s size can be precisely described. For each incoming frame, a probability distribution image of the target is created, where the target＇s area turns into a blob. The scale of this blob can be determined based on local maxima of differential scale-space filters. We employ the QP-TR trust region algorithm to search the local maxima of multi-scale normalized Laplacian filter of the probability distribution image to locate the target as well as determine its scale. In the second algorithm, we combine the template matching with the QP_TR method and achieved the real time performance. In the presented tracking examples, the two algorithms demonstrate their great improvement on tracking precision and runtime performance respectively.