一种高效的适用于复杂环境下物体跟踪的级联分类器

传统跟踪算法在复杂环境下容易发生漂移（drift）现象，本文改进了TLD跟踪技术算法提出了基于Sliding-window的局部搜索和全局搜索策略、积分直方图过滤器和随机Haar-like块特征过滤器。首先，采用积分直方图过滤器可以有效地过滤大量非目标子窗口块，从而减少后续过滤器特征匹配数；其次，利用随机Haar-like块特征过滤器能够解决跟踪算法在复杂环境（多物体、部分或较大区域遮挡、快速运动等）跟踪过程易发生漂移而导致跟踪精度的不足。本文结合TLD原始过滤器与新提出的两个过滤器组合而成的级联分类器，通过与主流的跟踪算法实验进行对比表明，级联分类器在稳定的背景或复杂环境的跟踪鲁棒性强、跟踪精度高，并且采用了局部和全局搜索策略提高了计算速度。

An Efficient Cascade Classifier for Object Tracking in Complex Conditions

This paper improves the TLD algorithm and proposes local and global search based on Sliding-window method , Integral Histogram Filter and Random Haar-like Feature Filter to solve the easily drift problem of the traditional tracking algorithm in complex conditions. First, we exploit the Integral Histogram Filter to reject the Sliding-window patches as quickly as possible to release the feature matching in the following filters. Then, we use Random Haar-like Feature Filter to overcome drift problem which causes the less accuracy during the object tracking in complex conditions (multi-object, partial or most occlusion, fast movement).We ultimately combine filters of the TLD algorithm and two new filters of our proposed. The experimental results show that the proposed approaches compared with the traditional tracking algorithms not only presents robust and tracking accuracy in stable background or complex conditions, but also obtains the best computing speed with the use of the local and global search.