| 基于DESO的Mean Shift目标跟踪算法研究 |
| |
| 引用本文： | 王晓卫,马晓军,周启煌. 基于DESO的Mean Shift目标跟踪算法研究[J]. 控制与决策, 2009, 24(2) |
| |
| 作者姓名： | 王晓卫 马晓军 周启煌 |
| |
| 作者单位： | 1. 陆军航空兵学院,机载设备系,北京,101123;装甲兵工程学院控制工程系,北京,100072
2. 装甲兵工程学院控制工程系,北京,100072 |
| |
| 基金项目： | 国家自然科学基金，武器装备预研基金 |
| |
| 摘 要： | 将基于DESO的运动预测算法和Mean Shift算法相结合,形成一种新的基于Mean Shift的快速目标跟踪算法.该算法以DESO预测位置作为Mean Shift算法下一帧候选模型的计算中心,实现了对快速运动目标的跟踪,并通过DESO对目标运动轨迹进行预测,较好地解决了目标完全遮挡时的跟踪问题.实验结果表明,该算法具有预测精度高、实时性好、抗遮挡能力强的优点.
|
| 关 键 词： | 目标跟踪 运动预测 均值迁移算法 微分扩张状态观测器 |
Study on Mean Shift tracking algorithm based on DESO |
| |
| WANG Xiao-wei,MA Xiao-jun,ZHOU Qi-huang. Study on Mean Shift tracking algorithm based on DESO[J]. Control and Decision, 2009, 24(2) |
| |
| Authors: | WANG Xiao-wei MA Xiao-jun ZHOU Qi-huang |
| |
| Affiliation: | 1.Department of Aviation Equipment;Army Aviation Institute;Beijing 101123;China;2.Department of Control Engineering;Academy of Armored Force Engineering;Beijing 100072;China |
| |
| Abstract: | The modified Mean Shift algorithm based differential extended state observer(DESO)is proposed by the combination of DESO and Mean Shift algorithm,which realizes tracking for target with high speed since it searches for target in the neighborhood of estimated position that DESO predicts.Moreover,the proposed algorithm uses target's motion information that DESO predicts,which can solve target occlusion preferably.Experiment results show the proposed algorithm has superior features,such as higher prediction pr... |
| |
| Keywords: | Object tracking Motion prediction Mean Shift algorithm Differential extended state observer |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
