Person tracking based on multi-template regression weighted mean shift

Song-min JIA; Lin-feng WEN; Li-jia WANG

doi:10.3788/OPE.20162409.2339

您当前的位置：

首页 >

文章列表页 >

Person tracking based on multi-template regression weighted mean shift

Information Sciences | 更新时间：2020-07-07

- Person tracking based on multi-template regression weighted mean shift
- Optics and Precision Engineering Vol. 24, Issue 9, Pages: 2339-2346(2016)
- 作者机构：
  
  1.北京工业大学信息学部, 北京 100124
  2.计算智能与智能系统北京市重点实验室, 北京 100124
  3.河北工业职业技术学院信息工程与自动化系, 河北石家庄 050091
- 作者简介：
- 基金信息：
- DOI：10.3788/OPE.20162409.2339
  CLC： TP391.4
- Received：11 April 2016，
  
  Accepted：20 June 2016，
  
  Published：2016-09
- 稿件说明：
移动端阅览
Song-min JIA, Lin-feng WEN, Li-jia WANG. Person tracking based on multi-template regression weighted mean shift[J]. Optics and precision engineering, 2016, 24(9): 2339-2346.
DOI：

Song-min JIA, Lin-feng WEN, Li-jia WANG. Person tracking based on multi-template regression weighted mean shift[J]. Optics and precision engineering, 2016, 24(9): 2339-2346. DOI： 10.3788/OPE.20162409.2339.

摘要

针对移动机器人跟踪人体目标时目标因角度大幅变化引起外观改变造成的跟踪无效，提出了多模板回归加权均值漂移跟踪方法。该方法通过建立目标的多模板模型，应用均值漂移算法实现目标跟踪。首先，根据前一帧均值漂移结果和当前帧头肩粗定位结果确定目标模板集，使其包含目标人体的位姿和角度改变。然后，采用多模板回归加权均值漂移实现目标的精确定位。在多模板均值漂移中引入回归模型实现颜色纹理特征与目标模型相似度之间的映射，从而控制模板数量，保证目标检测的实时性。最后，分别在视频图像和机器人目标跟踪平台上对所提方法进行实验验证。结果显示，图像处理平均时间为86.4 s/frame，满足机器人跟踪的实时性要求。该方法解决了目标特征在跟踪过程中发生变化的问题，提高了机器人跟踪时对目标人体特征变化的鲁棒性。

Abstract

To solve the invalid tracking of a human target caused by appearance variations due to large angle change of the target in a robot mobile tracking

a multi-template regression weighted mean-shift algorithm was proposed. The algorithm could implement the target tracking by building a multi-template model of the target and applying mean shift. Firstly

the template set was obtained according to the results from mean shift procedure of the last frame and the coarse location information of head-shoulder model of a current frame

by which the position and angle variation of the target person were included. Then

the multi-template regression weighted mean-shift algorithm was used to determine the precise location of the target person. The regression model was introduced to multi-template mean shift to implement a map from color-texture feature to the similarity of target model to limit the number of templates and to ensure the real-time performance of the target detection. Finally

the proposed algorithm was verified by videos and robot tracking tests. The results show that the image average treatment time is 86.4 ms/frame

which satisfies the requirement of person tracking for a mobile robot. The method solves the appearance variation problem of targets in tracking processing and improves the robustness of human targets to its feature variations.

关键词

Keywords

references

初红霞, 谢忠玉, 王君祥, 等.空间相关背景加权直方图均值漂移目标跟踪[J].控制与决策, 2014, 29(3):528-532.

CHU H X, XIE ZH Y, WANG J X, et al.. Mean shift target tracking with spatiogram corrected background-weighted histogram[J]. Control and Decision, 2014, 29(3):528-532. (in Chinese)

修春波, 魏世安.显著性直方图模型的Camshift跟踪方法[J].光学精密工程, 2015, 23(6):1749-1757.

XIU CH B, WEI SH A. Camshift tracking with saliency histogram[J]. Opt. Precision Eng., 2015, 23(6):1749-1757. (in Chinese)

周芳芳, 樊晓平, 叶榛.均值漂移算法的研究与应用[J].控制与决策, 2007, 22(8):841-847.

ZHOU F F, FAN X P, Y ZH. Mean shift research and applications[J]. Control and Decision, 2007, 22(8):841-847. (in Chinese)

郑超, 陈杰, 陶会峰, 等.基于改进亮度变化函数实现红外图像中行人跟踪[J].光学精密工程, 2015, 23(10):2980-2988.

ZHENG CH, CHEN J, TAO H F, et al.. Pedestrian tracking in FLIR imagery based on modified intensity variation function[J]. Opt. Precision Eng., 2015, 23(10):2980-2988. (in Chinese)

陈东成, 朱明, 高文, 等.在线加权多示例学习实时目标跟踪[J].光学精密工程, 2014, 22(6):1661-1667.

CHEN D CH, ZHU M, GAO W, et al.. Real-time object tracking via online weighted multiple instance learning[J]. Opt. Precision Eng., 2014, 22(6):1661-1667. (in Chinese)

苏娟, 杨小冈, 卢俊.鲁棒的前视红外目标多核跟踪算法[J].红外与激光工程, 2013, 42(2):543-548.

SU J, YANG X G, LU J. Robust multi-kernel tracking algorithm for FLIR target[J].Infrared and Laser Engineering, 2013, 42(2):543-548. (in Chinese)

李琦, 邵春福, 岳昊.核窗口尺寸和目标模型自适应的均值漂移跟踪[J].华南理工大学学报(自然科学版), 2013, 41(2):74-81.

LI Q, SHAO CH F, YUE H. Mean shift tracking with adaptive kernel window size and target model[J]. Journal of South China University of Technology(Natural Science Edition), 2013, 41(2):74-81. (in Chinese)

胡良梅, 段琳琳, 张旭东, 等.融合颜色信息与深度信息的运动目标检测方法[J].电子与信息学报, 2014, 36(9):2047-2052.

HU L M, DUAN L L, ZHANG X D, et al.. Moving object detection based on the fusion of color and depth information[J]. Journal of Electronics & Information Technology, 2014, 36(9):2047-2052. (in Chinese)

QIAO L Y, XU L X, GAO M. Infrared target tracking using bandwidth adaptive mean shift[J]. Infrared and Laser Engineering, 2015, 44(1):354-362.

王丽佳, 贾松敏, 李秀智, 等.分块多特征目标描述子的移动机器人目标跟踪[J].控制与决策, 2016, 31(2):337-342.

WANG L J, JIA S M, LI X ZH, et al.. Person tracking for robot using patches-based-multi-cues representation[J]. Control and Decision, 2016, 31(2):337-342. (in Chinese)

贾松敏, 王爽, 王丽佳, 等. CTF定位策略下基于多特征的智能机器人目标跟踪[J].上海交通大学学报, 2014, 48(7):1039-1052.

JIA S M, WANG SH, WANG L J, et al.. Human tracking based on multi-feature for intelligent robot under the CTF location strategy[J]. Journal of Shanghai Jiaotong University, 2014, 48(7):1039-1052. (in Chinese)

宋潇毅.基于纹理和颜色特征的图像检索[D].成都:电子科技大学, 2009. http://cdmd.cnki.com.cn/article/cdmd-10614-2009165749.htm

SONG X Y.Image retrieval based on texture and color[D].Chengchu:University of Electronic Science and Technology of China, 2009.(in Chinese)

张红颖, 胡正.融合局部三值数量和色度信息的均值漂移跟踪[J].电子与信息学报, 2014, 36(3):624-630.

ZHANG H Y, HU ZH. Mean shift tracking method combining local ternary number with hue information[J]. Journal of Electronics & Information Technology, 2014, 36(3):624-630. (in Chinese)

TURHAN M M, HANBAY D. Mean shift based object tracking supported with adaptive Kalman filter[C]. 23rd Signal Processing and Communications Applications Conference, 2015.

李冠彬, 吴贺丰.基于颜色纹理直方图的带权分块均值漂移目标跟踪算法[J].计算机辅助设计与图形学学报, 2011, 23(11):2059-2066.

LI G B, WU H F. Weighted fragments-based meanshift tracking using color-texture histogram[J]. Journal of Computer-Aided Design & Computer Graphics, 2011, 23(11):2059-2066. (in Chinese)

杨军, 高志升, 袁红照, 等.基于LBP特征和贝叶斯模型的单样本人脸识别[J].光电子.激光, 2011, 22(5):763-765.

YANG J, GAO ZH SH, YUAN H ZH, et al.. Single sample face recognition based on LBP feature and Bayes model[J]. Journal of Optoelectronics·Laser, 2011, 22(5):763-765. (in Chinese)

刘桂雄, 张晓平, 周松斌.基于最小二乘支持向量回归机的无线传感器网络目标定位法[J].光学精密工程, 2009, 17(7):1766-1773.

LIU G X, ZHANG X P, ZHOU S B. Target localization in wireless sensor networks based on LSSVR[J]. Opt. Precision Eng., 2009, 17(7):1766-1773. (in Chinese)

汪奇生.线性回归模型的总体最小二乘平差算法及其应用研究[D].昆明:昆明理工大学, 2014. http://cdmd.cnki.com.cn/article/cdmd-10674-1015547948.htm

WANG Q SH.Research of total least squares adjustment algorithm based on linear regression model and its application[D].Kunming:Kunming University of Science and Technology, 2014.(in Chinese)

汪方斌, 洪津, 孙晓兵, 等.基于支持向量机预测偏振模式[J].光学精密工程, 2014, 22(11):2914-2922.

WANG F B, HONG J, SUN X B, et al.. Prediction of polarization pattern by SVM[J]. Opt. Precision Eng., 2014, 22(11):2914-2922. (in Chinese)

ZHANG T J, BAI X F, HAN Q. Contour detection based on support vector regression in natural images[J]. International Journal of Digital Content Technology and its Applications, 2012, 6(22):713-721.

Views

433

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Moving object detection based on variant block difference in complex scenes

Real-time restoration of motion-blurred video images on GPU

Related Author

ZHU Juan-juan

GUO Bao-long

WANG Jing

LI Shi

Related Institution

西安电子科技大学机电工程学院

中国科学院长春光学精密机械与物理研究所,吉林长春130033

杭州电子科技大学图形图像研究所

AI问答

Address：No.3888 Dong Nanhu Road, Changchun, Jilin, China Postal code：130033
Tel：0431-86176855 Email：gxjmgc@ciomp.ac.cn
Technical support is provided by Beijing Founder electronics co., LTD 吉ICP备11002662号-17 京公网安备11010802024621
It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.

⁰