这是一种新的单目标长时间（long term tracking） 跟踪算法。该算法与传统跟踪算法的显著区别在于将传统的跟踪算法和传统的检测算法相结合来解决被跟踪目标在被跟踪过程中发生的形变、部分遮挡等问题。同时，通过一种改进的在线学习机制不断更新跟踪模块的显著特征点和检测模块的目标模型及相关参数，从而使得跟踪效果更加稳定、鲁棒、可靠。-This is a new single target for a long time (long term tracking) tracking algorithm. The algorithm and the conventional tracking algorithm significant difference is that the conventional tracking algorithms and conventional detection algorithm combines a deformation occurs to solve the target being tracked in the tracking process, partial occlusion and other issues. Meanwhile, through an improved online learning mechanism constantly updated tracking module significantly the target model and related parameters of the feature points and the detection module, so that the tracking results more stable, robust and reliable.