一种新的基于Mean Shift的目标三自由度跟踪算法
doi: 10.3724/SP.J.1146.2006.01705
A New Mean Shift Based Algorithm for Tracking Targets with Three Degrees of Freedom
-
摘要: 标准Mean Shift跟踪算法仅能确定目标形心位置，而不能确定其旋转角，在跟踪细长形目标时鲁棒性不好。为此，该文提出了一种三自由度Mean Shift跟踪算法，新算法在计算目标特征分布直方图时，用像素的位置转角及其到目标形心的归一化距离加权，并将像素在局部坐标系下的特征转角作为新特征引入。这种新的目标表示模型能够方便地纳入Mean Shift优化框架，通过迭代求解，可同时精确确定目标的形心位置和方位指向。实验结果表明该算法精度高，计算量小。
-
关键词:
- 目标跟踪; 均值漂移; 转角定位
Abstract: Standard Mean Shift tracker can only successfully locate the object center, but fail to find its orientation, which make it not robust to track thin object. To remedy this, an improved mean shift tracker is proposed in this paper. The new tracker use new object representation, where pixels are weighted with both their position-angles and normalized distances from target center, furthermore, pixels feature-angle, which can be seen as new feature, is introduced in. The new object representation can be conveniently integrated into the optimization framework of mean shift. By iterative optimization, both the location and orientation of targets can be precisely determined. Experimental results show the algorithm can get precise tracking results with low computational cost. -
Comaniciu D, Ramesh V, and Meer P. Kernel-based objecttracking[J].IEEE Trans. on Pattern Analysis and MachineIntelligence.2003, 25(5):564-577[2]Zhao Q and Tao H. Object tracking using color correlogram.The 2nd Joint IEEE International Workshop on VisualSurveillance and Performance Evaluation of Tracking andSurveillance, Beijing, 2005: 263-270.[3]Bradski G R. Computer vision face tracking for use in aperceptual user interface. IEEE Workshop on Applications ofComputer Vision, Princeton, 1998: 214-219.[4]贾静平, 柴艳妹, 赵荣椿. 一种健壮的目标多自由度MeanShift 序列图像跟踪算法. 中国图象图形学报, 2006, 11(5):707-713.Jia J-P, Chai Y-M, and Zhao R-C. Robust tracking of objectsin image sequences using multiple degrees of freedom meanshift algorithm. Chinese Journal of Image and Graphics, 2006,11(5): 707-713.[5]Yang C J, Duraiswami R, and Davis L. Efficient spatialfeaturetracking via the mean-shift and a new similaritymeasure. Proceeding of IEEE Conference on ComputerVision and Pattern Recognition, San Diego, USA, 2005:176-183.Zhang H H, Huang W M, and Huang Z Y, et al.. Affine objecttracking with kernel-based spatial-color representation. IEEEConference on Computer Vision and Pattern Recognition,San Diego, 2005: 293-300.[6]Comaniciu D and Meer P. Mean shift: A robust approachtoward feature space analysis[J].IEEE Trans. on PatternAnalysis and Machine Intelligence.2002, 24(5):603-619 -
计量
- 文章访问数: 2880
- HTML全文浏览量: 64
- PDF下载量: 1139
- 被引次数: 0
下载:
