基于TWH与FB-error约束的均值漂移无人机跟踪算法  

Mean-shift UAV tracking algorithm design based on TWH and FB-error constraints

作　　者：李一波[1] 张琦[1] 

机构地区：[1]沈阳航空航天大学自动化学院,沈阳110136

出　　处：《沈阳航空航天大学学报》2016年第5期49-54,共6页Journal of Shenyang Aerospace University

基　　金：辽宁省高等学校优秀人才支持计划(项目编号:LJQ2014018);辽宁省教育厅一般项目(项目编号:L2014066)

摘　　要：针对无人机跟踪过程中目标遮挡和目标背景变化等因素导致的跟踪失败现象,提出一种M TF(M ean-shift by TWH and FB-error)跟踪算法。首先,在M ean-shift跟踪框架下引入目标加权直方图(TWH:Target-Weighted Histogram)描述目标,即在跟踪过程中,用目标的局部背景来削弱所有区域的内部背景特征,使目标特征突出;其次,添加FB-error约束,在目标被部分遮挡时,通过使用FB-error相关加权函数把目标当前位置的预测结果与Mean-shift矢量计算出的位置结果联合起来估计目标在t时刻的最终位置。实验表明,此跟踪算法在跟踪精度上有较大突破。For the tracking failures caused by the target occlusion and target background changing during UAV tracking, a tracking algorithm named mean-shift by TWH and FB-error（MTF） was proposed. First, the target-weighted histogram（TWH） was introduced in the mean-shift tracking framework to describe the tar- get. That was in the tracking process, the internal background features of all regions were weakened by the local background of the target to highlight the target feature. Second, the FB-error constraint was added. When the target was partially obscured, by using the weighted function of FB-error, the predicted result of the current target position and the calculated result by mean-shift vector were combined to estimate the final target position at time t. The experimental results show that the proposed tracking algorithm has a great improvement in tracking accuracy.

关 键 词：UAV 跟踪 均值漂移算法 向前向后误差算法 

分 类 号：V211[航空宇航科学与技术—航空宇航推进理论与工程]