针对目标跟踪容易受到遮挡、形变和光照变化影响的问题, 在粒子滤波框架下提出一种基于多特征和局部联合稀疏表示的目标跟踪算法。利用HSV空间建立目标的颜色表观模型; 利用增强的中心对称局部二值模式建立目标的纹理表观模型, 并用局部联合稀疏编码表示。综合颜色和纹理特征计算候选区域与目标的相似性, 并利用最大后验概率估计目标当前状态。每2帧判断一次目标表观模型是否需要更新, 减少了因频繁更新目标造成的累积误差。利用visual tracker benchmark数据集与其他4种跟踪算法进行了对比实验, 结果表明, 本文算法的整体精确度和成功率分别为83.5%和79.6%。本文算法在存在遮挡、形变和光照变化的情况下, 能够准确稳定地跟踪目标。

Aimed at the problem of occlusion, deformation and illumination in the object tracking, an object tracking method based on multi-feature and local joint sparse representation is proposed within particle filter framework. The color model of the object is established by using HSV space. The texture apparent model of the object is established by using the enhanced center symmetric local binary patterns and represented by the local joint sparse coding. Integrating the color and texture features,the similarities of the object and candidate regions are computed. The object state is estimated by the maximum posterior probability. Whether the object model need to be updated is judged every two frames, which reduces the accumulative errors caused by frequent updates. The proposed method is compared with the other four methods by using visual tracker benchmark data set. Experimental results show that the overall accuracy and success rate of the proposed method is 83.5% and 79.6% respectively. In the case of occlusion, deformation and illumination, the proposed method can track the object accurately and steadily.