针对现有Mean shift跟踪算法在目标被遮挡、跟踪场景变化时,跟踪误差变大甚至丢失目标的问题,提出了一种基于有效特征筛选的Mean shift运动目标跟踪算法.首先通过对目标特征的优化筛选,改善了现有Mean shift算法因目标特征多而造成计算时间较长,在目标发生较大变化时跟踪精度降低的情况.更能有效地表征目标特征,减少跟踪误差,增强特征集对目标的描述能力.同时给出目标模板更新的方法,在目标发生明显变化时,能自适应地更新特征集,进一步提高跟踪精度.仿真结果表明:文中方法具有更好的跟踪精度,计算时间较小,对遮挡、场景变化有更好的鲁棒性.%Aim. The introduction of the full paper reviews a number of papers in the open literature and then proposes what we believe to be a new and effective algorithm, which is explained in sections 1 and 2. Section 1 briefs the traditional Mean shift algorithm. The core of section 2 consists of: (1) our new and effective feature extraction algorithm improves the present Mean shift tracking algorithms in that the calculation time is shorter and that the tracking accuracy is higher; the target' s features can be displayed more effectively, tracking errors can be reduced , and the description of the features set can be enhanced; eqs. (10) and (11) are worth particular attention; (2) the target updating method we propose can adaptively update feature set when the target changes greatly and suddenly, and enhance tracking accuracy; eqs. ( 12) , ( 13) , and (14) are worth particular attention. Simulation results, presented in Figs. 1 through 4, and their analysis show preliminarily that our new effective feature extraction algorithm has indeed higher localization precision and requires indeed less computational time.
展开▼
