目标尺度自适应的Mean Shift跟踪算法

兵工学报 ›› 2011, Vol. 32 ›› Issue (2): 210-216.

目标尺度自适应的Mean Shift跟踪算法

康一梅¹，谢晚冬¹，胡江²，黄琪¹

(1．北京航空航天大学软件学院，北京 100037;2．中国兵器工业计算机应用技术研究所，北京 100089)

收稿日期:2009-09-07 修回日期:2009-09-07 上线日期:2014-05-04
通讯作者: 康一梅 E-mail:kangyimei@buaa.edu.cn
作者简介:康一梅（1968—），女，博士，副教授

Target Scale Adaptive Mean Shift Tracking Algorithm

KANG Yi-mei¹, XIE Wan-dong¹, HU Jiang², HUANG Qi¹

(1.Software School, Beihang University, Beijing 100037, China;2.Beijing Institute of Computer Application, Beijing 100089, China)

Received:2009-09-07 Revised:2009-09-07 Online:2014-05-04
Contact: KANG Yi-mei E-mail:kangyimei@buaa.edu.cn

摘要/Abstract

摘要： 传统的Mean Shift跟踪算法由于固定了核函数的带宽，因而不能很好地对图像尺度不断变换的目标进行有效的跟踪。针对这一不足，提出了一种将尺度变换和Mean Shift跟踪相结合的目标跟踪方法。在每一帧中，先由 Mean Shift得到目标位置，然后利用仿射变换原理计算得到相邻两帧之间目标的仿射变换矩阵，并利用该矩阵对目标位置和大小进行修正。实验表明改进算法有效的提高了Mean Shift算法在目标尺度变化时的跟踪稳定性，对目标的尺度变化具有适应性。

关键词: 人工智能, Mean Shift, 目标跟踪, 仿射变换, 尺度自适应

Abstract: The traditional Mean-Shift tracking algorithm is not applied to track a size-changing target effectively due to the fixed band-width of its kernel function. For this reason, a new algorithm is proposed. In each frame, Mean Shift tracking algorithm is employed to get the target location, and then the affine structure between frames is calculated to re-correct the position and size of the target. Different experiments prove that the present method has better stability and robustness than the traditional algorithm.

Key words: artificial in telligence, mean shift, target tracking, affine structure, scale adaptation

中图分类号:

TP391

康一梅，谢晚冬，胡江，黄琪. 目标尺度自适应的Mean Shift跟踪算法[J]. 兵工学报, 2011, 32(2): 210-216.

KANG Yi-mei, XIE Wan-dong, HU Jiang, HUANG Qi. Target Scale Adaptive Mean Shift Tracking Algorithm[J]. Acta Armamentarii, 2011, 32(2): 210-216.

参考文献

［1］ Fukunaga K, Hostetler L D.The estimation of the gradient of a density function with applications in pattern recognition[J]. IEEE Trans. on Information Theory, 1975,21(1):32-40.
［2］ CHENG Y Z. Mean Shift, mode seeking and clustering[J]. IEEE Trans. on Pattern Analysis and Machine Intelligence,1995,17(8):790-799.
［3］ Comaniciu D, Meer P.Mean Shift analysis and applications[C]∥Proceedings of the 7^th IEEE International Conference. Computer Vision, Kerkyra. 1999.
［4］ Comaniciu D, Meer P.Mean Shift: a robust application toward feature space analysis[J]. IEEE Trans on Pattern Analysis and Machine Intelligence, 2002,24(5):603-619.
［5］ Comaniciu D, Meer P.Robust analysis of feature spaces: color image segmentation[C]∥Proceedings of 1997 IEEE Conference. Computer Vision and Pattern Recognition, San Juan. 1997.
［6］ Comaniciu D, Ramesh V, Meer P.Kernel-Based object tracking[J]. IEEE Trans on Pattern Analysis and Machine Intelligence, 2003,25(5):564-575.
［7］ Collins R T.Mean Shift blob tracking through scale space[J]. IEEE International Conference on Computer Vision and Pattern Recognition, Victor Graphics, 2003: 234-240.
［8］ Bradski G R.Computer vision face tracking for use in a perceptual user interface[J]. IEEE Workshop on Applications of Computer Vision, Stoughton: Printing House, 1998: 214-219.
［9］彭宁嵩，杨杰，刘志，等. Mean Shift跟踪算法中核函数带宽的自动选取[J].软件学报，2005,16(9):1542-1550.
PENG Ning-song, YANG Jie, LIU Zhi, et al. Automatic selection of bandwidth for mean-shift object tracking［J］. Journal of Software, 2005, 16(9): 1542-1550. (in Chinese)
［10］康文静，丁雪梅，刘功亮，等. 带宽自适应Mean Shift跟踪算法[J].光电子·激光，2008,(1)：135-138.
KANG Wen-jing, DING Xue-mei, LIU Gong-liang, et al. Bandwidth adaptive mean-shift tracking algorithm［J］. Journal of Optoelectronics Laser, 2008,(1): 135-138. (in Chinese)
［11］钱惠敏，茅耀斌，王执铨.自动选择跟踪窗尺度的Mean Shift算法[J].中国图象图形学报，2007，2(2):246-249.
QIAN Hui-min, MAO Yao-bin, WANG Zhi-quan. Mean shift tracking with self-updatingtracking window［J］. Journal of Image and Graphics, 2007,2(2):246-249. (in Chinese)
［12］覃剑，曾孝平，李勇明.基于边界力的Mean Shift核窗宽自适应算法[J].软件学报，2009，20（7）：1726-1734.
QIN Jian, ZENG Xiao-ping, LI Yong-ming. Algorithm of adaptive kernel-bandwidth for mean-shift based on boundary force［J］. Journal of Software, 2009,20(7):1726-1734. (in Chinese)
［13］朱胜利，朱善安.核函数带宽自适应的Mean Shift目标跟踪算法[J].光电工程，2006，33(8)：11-16.
ZHU Sheng-li, ZHU Shan-an. Algorithm of target tracking based on mean shift with adaptive bandwidth of kernel funciton［J］. Opto-Electronic Engineeirng, 2006,33(8):11-16. (in Chinese)
［14］ Fischler M A, Bolles R C. Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography[J]. Comm. ACM, 1981,24(6):381-395.

[1]	张昀普，单甘霖. 道路约束下多传感器协同地面目标跟踪的管理方法[J]. 兵工学报, 2022, 43(3): 542-555.
[2]	李纪三，侯娇，刘溶，任渊. 基于贝叶斯滤波算法的警戒相控阵雷达目标跟踪时间资源优化分配算法[J]. 兵工学报, 2021, 42(9): 1902-1910.
[3]	姜珊，底晓强，韩成. 融合时空特性的孪生网络视觉跟踪[J]. 兵工学报, 2021, 42(9): 1940-1950.
[4]	丁海婷，周琳，刁伟峰. 基于背包问题的多相控阵雷达多目标跟踪时间资源管理算法[J]. 兵工学报, 2021, 42(5): 997-1003.
[5]	张永梅，赖裕平，马健喆，冯超，束颉. 基于视频的装甲车和飞机检测跟踪及轨迹预测算法[J]. 兵工学报, 2021, 42(3): 545-554.
[6]	张君彪，熊家军，兰旭辉，李凡，陈新，席秋实. 基于自适应多通道交互式多模型的高超声速滑翔飞行器跟踪方法[J]. 兵工学报, 2021, 42(12): 2626-2636.
[7]	邹汝平，刘建书. 基于概率假设密度滤波与无迹Kalman滤波的多目标跟踪与识别[J]. 兵工学报, 2020, 41(8): 1502-1508.
[8]	卢莉萍，郑潇. 基于旋转反射镜光学成像的动态目标跟踪方法[J]. 兵工学报, 2020, 41(8): 1558-1565.
[9]	邵鹏飞，王蕾，王方勇. 基于序贯蒙特卡洛与概率假设密度滤波的主动分布式声纳多目标跟踪[J]. 兵工学报, 2020, 41(5): 941-949.
[10]	张伽伟，喻鹏，姜润翔, 孙宝全. 基于舰船电场的目标跟踪方法研究[J]. 兵工学报, 2020, 41(3): 559-566.
[11]	杨鑫, 许卫东, 贾其. 一种基于特征统计的动态伪装效果评估方法[J]. 兵工学报, 2019, 40(8): 1693-1699.
[12]	刘昊，张策，丁文韬. 基于智能对抗进化的联合火力打击任务规划方法[J]. 兵工学报, 2019, 40(6): 1287-1296.
[13]	马艳，刘小东. 状态自适应无迹卡尔曼滤波算法及其在水下机动目标跟踪中的应用[J]. 兵工学报, 2019, 40(2): 361-368.
[14]	张蛟，邢士勇，王建华，李茂，陆建峰. 一种具有角速度量测的多站无源目标跟踪方法[J]. 兵工学报, 2019, 40(1): 107-114.
[15]	曾溢良，蓝金辉，邹金霖. 滑动置信度约束的红外弱小目标跟踪算法研究[J]. 兵工学报, 2017, 38(9): 1771-1778.