中文摘要：

针对中小型UAV（Unmanned Aerial Vehicle）目标跟踪过程中常见的目标跟踪实时性不高,以及目标被遮挡易导致目标丢失问题展开研究。对比了目前现有的视频目标跟踪算法的优缺点,提出了一种M_APF算法,该算法是一种以随机性跟踪算法作为算法的主体框架,将确定性跟踪算法嵌入其中来降低跟踪时所需的样本数,提高算法的实时性,将其与Meanshift算法,APF算法等跟踪算法进行对比研究,仿真结果表明：M_APF算法在对UAV平台运动目标跟踪过程中实时性及鲁棒性优于Meanshift和APF等算法。

英文摘要：

Research on the problems of real time target tracking and target occlusion in the UAV target tracking system,the development of the UAV target tracking methods is discussed.An M-APF Algorithm is proposed,which adopts the APF as the main framework of the tracking algorithm,then the Meanshift is applied to calculate the offset of a few auxiliary particle and move them to the local optimum position of the observed values.Simulation results show that the calculation cost and robustness of the M-APF is superior to Meanshift and APF algorithm,and satisfied the UAV target tracking requirements.