局部感知下的稀疏优化目标跟踪方法

针对传统稀疏表示跟踪算法在复杂背景中易出现跟踪漂移问题，该文提出一种局部感知下的稀疏优化目标跟踪方法。首先，将首帧确定的目标区域进行非重叠均匀分割，并利用目标的全局特征和局部特征联合建模。然后，提出一种局部感知校验方法约束稀疏优化匹配过程，从而确定最优匹配样本。最后，在模板更新中提出一种决策方法对遮挡进行检测，并针对不同遮挡情况采取相应的更新策略，使得更新后的模板集更加完善。实验在10个标准库视频序列中测试，并与目前较流行的目标跟踪算法在跟踪效果、成功率等方面进行比较，实验结果表明，提出的跟踪方法在局部遮挡、目标形变、复杂背景等条件下跟踪准确、适应性强。

Object Tracking Method Based on Sparse Optimization of Local Sensing

The problem of tracking drift is produced easily by traditional sparse representation tracking methods in complex scene. To solve this problem, a novel tracking approach based on sparse optimization of local sensing is proposed. Firstly, the object area of the first frame is divided into non-overlapping uniform segmentation, and building the template set using global features and local features. Then, a local sensing correction method for constraining sparse optimization matching process is utilized to determine the optimal matching samples. Finally, a new method of occlusion decision is used to detect occlusion, and updating strategies are adopted according to different occlusion conditions, which makes the template sets more complete in the process of template update. The experiments compare with state-of-the-art tracking algorithms on 10 tracking test sequences of benchmark library. Experiment results indicate that the proposed method possesses characteristics of accurate tracking and strong adaptability in the conditions of partial occlusion, deformation, and complex background.