
浏览全部资源
扫码关注微信
1. 中国科学院 长春光学精密机械与物理研究所,吉林 长春,中国,130033
2. 中国科学院 研究生院 北京,100049
收稿日期:2011-12-05,
修回日期:2012-02-21,
网络出版日期:2012-06-15,
纸质出版日期:2012-06-15
移动端阅览
王田, 刘伟宁, 韩广良, 杜超, 刘恋. 基于改进MeanShift的目标跟踪算法[J]. 液晶与显示, 2012,(3): 396-400
WANG Tian, LIU Wei-ning, HAN Guang-liang, DU Chao, LIU Lian. Target Tracking Algorithm Based on Improved Meanshift[J]. , 2012,(3): 396-400
王田, 刘伟宁, 韩广良, 杜超, 刘恋. 基于改进MeanShift的目标跟踪算法[J]. 液晶与显示, 2012,(3): 396-400 DOI: 10.3788/YJYXS20122703.0396.
WANG Tian, LIU Wei-ning, HAN Guang-liang, DU Chao, LIU Lian. Target Tracking Algorithm Based on Improved Meanshift[J]. , 2012,(3): 396-400 DOI: 10.3788/YJYXS20122703.0396.
针对传统Meanshift算法在某些干扰或遮挡情况下不能保证跟踪的准确性
以及目标模型内的背景像素也会造成定位偏差的问题
提出一种基于MeanShift的改进算法。首先对目标模型进行改进
通过目标与背景的区分度引入权系数
在目标模型中进行加权处理
可达到降低目标模型内背景像素对跟踪定位精度的影响。然后
将跟踪窗进行分块
对各子块使用改进目标模型的Meanshift算法进行跟踪。最后
用匹配度最大的两个子块加权决定目标的最终位置
从而在目标发生遮挡时能有效剔除被遮挡子块对目标定位的影响。实验表明
在复杂背景下
新算法仍然可以有效、准确地跟踪运动目标。
Aiming at the problem that traditional Meanshift algorithm can not guarantee tracking accuracy in certain interference or occlusion case and background pixel in object model will induce localization error
a improved targets tracking algorithm based on Meanshift algorithm was proposed. Firstly
the object model is improved by introducing weights which is decided by divisional degree between object and background
and then the weight was used in object model to reduce the localization error of object tracking. Secondly
the target in the tracking window is divided into a number of fragments
the improved Mean Shift algorithm is used in every fragments separately. Finally
the match degree of each fragment is computed and the fragment with the highest match degree is involved to achieve the whole targets coordinates and to avoid the influence of occluded fragments on the object location. Experimental results show that the new method can effectively and accurately track moving target in the cluttering background.
Chen Y Z. Mean shift, mode seeking, and clustering [J].IEEE Transactions on Pattern Analysis Machine Intelligence, 1995, 17(8): 790-799.[2] Comanicu D, Meer P. Mean shift: A robust approach toward feature space analysis [J].IEEE Transactions on Pattern Analysis and Machine Intelligence, 2002, 24(5): 603-619.[3] Comaniciu D, Ramesh V, Meer P. Real-time tracking of non-rigid objects using mean shift //Proc. of the IEEE Conf. on Computer Vision and Pattern Recognition (CVPR),Hiton Head Island. USA:IEEE Computer Society, 2000:142-149.[4] 孙中森,孙俊喜,宋建中,等.一种抗遮挡的运动目标跟踪算法 [J]. 光学 精密工程, 2007,15(2):267-171.[5] 韩晓波.基于背景建模和动态分块的目标跟踪 [J]. 电子技术研发, 2010, 37(10):21-23.[6] Magglo E, Cavallaro A. Multi-part target representation for color tracking //IEEE International Conference on Image Processing, Genova, Italy:IEEE,2005:729-732.[7] 颜佳,吴敏渊,陈淑珍,等.应用Mean Shift和分块的抗遮挡跟踪 [J]. 光学 精密工程, 2010,18(6):1413-1419.[8] 刘扬,张云峰,董月芳. 复杂背景下抗遮挡的运动目标跟踪算法 [J]. 液晶与显示,2010,25(6):890-895.[9] Collins R T, Liu Y,Leordeanu M. Online selection of discriminative tracking features [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2005,27(10):1631-1643.[10] 刘翔,周桢.基于分块背景建模的运动目标检测技术 [J]. 液晶与显示,2011,26(6):831-835.
0
浏览量
75
下载量
13
CSCD
关联资源
相关文章
相关作者
相关机构
京公网安备11010802024621