基于局部背景加权直方图的目标跟踪

Object tracking based on local background weighted histogram

针对传统的颜色直方图和加权颜色直方图跟踪算法难以在复杂环境下对目标进行有效跟踪的问题,提出了一种基于局部背景加权直方图的目标跟踪算法,该算法将目标的局部背景看作上下文,将其引入目标表征。在粒子滤波跟踪算法框架下,用局部背景加权直方图来表征目标,以增强目标与背景的鉴别性,从而突出目标区域内的前景信息。各种场景的实验结果比较表明,提出的跟踪算法比传统的颜色直方图和加权颜色直方图跟踪算法具有更好的稳定性和鲁棒性,特别是针对目标被局部遮挡及跟踪环境光照变化较大等情况。

For the object tracking based on traditional color histogram and weighted color histogram may be lost in the complex environment.An improved object tracking based on the local background weighted histogram（LBWH）is approached.The proposed tracking algorithm treats local background as the context,and introduces it into target representation.As a result,a target description,the LBWH is proposed in this paper.The LBWH enhances the discrimination between the target and background,so that highlights the foreground in the target region.An extensive number of comparative experiments show that the proposed tracking algorithm is more stable and robust than the traditional color histogram and weighted color histogram tracking algorithms,especially in the case of the object partial occlusion and illumination variation.