基于均值漂移的运动目标跟踪算法研究
Research of moving target tracking algorithm based on Mean-Shift
-
摘要: 为了实现在复杂环境下运动目标跟踪,提出了一种基于均值漂移的运动目标跟踪算法.该算法用帧间差法在复杂背景中获取目标模型,引入卡尔曼滤波器减少均值漂移算法迭代次数、解决遮挡问题,通过调节核函数带宽来改变跟踪窗口大小,以保证获取目标完整信息.实验结果表明,该算法不需要目标的先验知识即能实现复杂背景下对运动目标稳定、准确的跟踪,对目标遮挡有很好的鲁棒性.Abstract: A moving object tracking algorithm based on the Mean-Shift algorithm was proposed for the accurately tracking of moving target under complex environments. This algorithm obtained the target model through the frame difference method;then through Kalman filter to reduce the iteration times of Mean-Shift algorithm and solve occlusion;lastly through adjusting the kernel function bandwidth to change the size of tracking window in order to access the complete information of moving target. Experiment results showed that the algorithm can track the moving target stably and accurately under complex environment and robustness to occlusion without the prior information.
-
Key words:
- Mean-Shift algorithm /
- Kalman filter /
- target tracking /
- adaptive window /
- object occlusion
-
-
[1]
Comaniciu D, Ramesh V, Meer P. Real-time tracking of non-rigid objects using Mean shift[C]//Proc IEEE Conference on Computer Vision and Pattern Recognition,Hilton Head Island:SC,2000:142-149.
-
[2]
曹玉华,吴小俊,段先华,等. 基于背景提取和扩展均值漂移算法的目标跟踪[J].计算机工程与应用, 2009,45(13):194.
-
[3]
王明佳,武治国,韩广良,等. 自适应窗口选取的MeanShift目标跟踪[J].光子学报,2012,41(1):67.
-
[4]
郝志成,朱明,刘微. 复杂背景下目标的快速提取与跟踪[J].吉林大学学报:工学版,2006,36(2):259.
-
[5]
顾裕丰,刘国栋. 一种复杂场景下的运动目标跟踪算法[J].计算机技术及应用,2012,38(1):122.
-
[6]
刘继艳,潘建寿,吴亚鹏,等. 结合Kalman滤波器的Mean-Shift跟踪算法[J].计算机工程与应用, 2009, 45(12):184.
-
[7]
王娟,刘教民,孟军英,等. 改进均值漂移算法在运动目标跟踪中的研究[J].系统仿真学报,2012,24(9):1896.
-
[8]
欧阳宁,游锦海,莫建文,等. 遮挡情况下多目标跟踪算法[J].计算机应用研究,2010,27(5):1984.
-
[1]
-
点击查看大图
计量
- PDF下载量: 60
- 文章访问数: 966
- 引证文献数: 0
下载:
