[1]李晓峰,邢金明.融合时空多特征表示的运动人体目标跟踪算法[J].应用科技,2020,47(4):26-31,41.[doi:10.11991/yykj.202005007]
 LI Xiaofeng,XING Jinming.Tracking algorithm of a moving human body target using multi-feature representation of fused time and space[J].Applied science and technology,2020,47(4):26-31,41.[doi:10.11991/yykj.202005007]
点击复制

融合时空多特征表示的运动人体目标跟踪算法(/HTML)
分享到:

《应用科技》[ISSN:1009-671X/CN:23-1191/U]

卷:
第47卷
期数:
2020年4期
页码:
26-31,41
栏目:
智能科学与技术
出版日期:
2020-07-05

文章信息/Info

Title:
Tracking algorithm of a moving human body target using multi-feature representation of fused time and space
作者:
李晓峰1 邢金明2
1. 黑龙江外国语学院 信息工程系,黑龙江 哈尔滨 150025;
2. 东北师范大学 体育学院,吉林 长春 130024
Author(s):
LI Xiaofeng1 XING Jinming2
1. Department of Information Engineering, Heilongjiang International University, Harbin 150025, China;
2. School of Physical Education, Northeast Normal University, Changchun 130024, China
关键词:
融合时空多特征表示运动人体目标图像跟踪算法检测像素点跟踪耗时运动速度
Keywords:
fused time and spacemulti-feature representationmoving human bodytarget imagetracking algorithmdetection of pixelstime consumption of trackingmovement speed
分类号:
TP391
DOI:
10.11991/yykj.202005007
文献标志码:
A
摘要:
为提升运动人体目标的跟踪效果,缩短目标跟踪耗时,提出融合时空多特征表示的运动人体目标跟踪算法。利用运动人体目标位置的获取时间关系确定目标初始运动速度,根据目标区域的质心位置计算搜索窗,提取运动人体目标位置;融合待检测像素点、像素点矢量及最大似然估计值3大特征,将融合多特征表示引入运动人体目标的联合概率密度函数,利用运动人体目标检测门限检测运动人体目标图像像素点,确定运动人体目标区域;通过对运动人体目标的重采样及状态转移,完成运动人体目标的跟踪。实验结果表明:所提运动人体目标跟踪算法的跟踪准确率高到92%左右,跟踪耗时较短、跟踪查全率较好,跟踪效果得到了提升。
Abstract:
In order to improve the tracking effect of a moving human body target and reduce the time consumption of target tracking, we propose a moving human body target tracking algorithm by fusing spatiotemporal multi-feature representations. Using the acquisition time relationship of the target position of the moving human body to determine the initial speed of the target, and calculate the search window based on the position of centroid of the target imaging area, and extract the target position of the moving human body; the multi-feature representation is introduced to the joint probability density function of the moving human target, and the moving human target detection threshold is used to detect the imaging pixels to determine the moving human target area. After resampling and state-transferring, the tracking of the moving human body target is completed. The experimental results show that the tracking accuracy of the proposed algorithm is about 92%, with shorter tracking time and better tracking recall rate, improving the tracking effect.

参考文献/References:

[1] 董潇潇, 何小海, 吴晓红, 等. 基于注意力掩模融合的目标检测算法[J]. 液晶与显示, 2019, 34(8): 825–833
[2] CHEN Xiao, LI Yaan, LI Yuxing, et al. Active sonar target tracking based on the GM-CPHD filter algorithm[J]. Journal of Northwestern Polytechnical University, 2018, 36(4): 656–663.
[3] 程轩, 宋骊平, 姬红兵. 基于演化网络模型的箱粒子CPHD群目标跟踪[J]. 系统工程与电子技术, 2018, 40(5): 961–967
[4] 陈超. 高斯混合模型结合加权似然的目标跟踪算法[J]. 计算机工程与应用, 2019(12): 124–131
[5] 石曼曼, 李玲. 基于图像序列的人体运动跟踪算法[J]. 现代电子技术, 2017, 40(9): 25–28
[6] 孙彦景, 石韫开, 云霄, 等. 基于多层卷积特征的自适应决策融合目标跟踪算法[J]. 电子与信息学报, 2019, 41(10): 2464–2470
[7] SINGH P R, WANG Y, CHARGE P. Bistatic MIMO radar for near field source localization using PARAFAC[J]. Electronics letters, 2016, 52(12): 1060–1061.
[8] WANG Jiadong. ZHANG Lei, DU Lan, et al. Optimal coherent processing interval selection for aerial maneuvering target imaging using tracking information[J]. IEEE sensors journal, 2018, 18(10): 4117–4128.
[9] CHEN Chaoxiang, LI Wenshu, YE Shiping, et al. Real-time tracking algorithm for moving target based on dynamic monitoring video[J]. International journal of multimedia and ubiquitous engineering, 2016, 11(9): 157–164.
[10] 王春峰, 苏荔, 黄庆明. 基于卷积神经网络的时空融合的无参考视频质量评价方法[J]. 中国科学院大学学报, 2018, 35(4): 544–549
[11] 刘德喜, 夏先益, 万常选, 等. 基于多特征融合的在线论坛用户心理健康自动评估[J]. 计算机学报, 2019, 42(7): 1553–1569
[12] YUAN J, LI Y, DONG X, et al. Target tracking with a mobile robot based on on-line classification for motion patterns[J]. Chinese journal of scientific instrument, 2017, 38(3): 568–577.
[13] HAN Gao, ZHANG Yanmei, GUO Haichao. A compressive sensing algorithm using truncated SVD for three-dimensional laser imaging of space-continuous targets[J]. Journal of modern optics, 2016, 63(21): 2166–2172.
[14] UNGI T, ABOLMAESUMI P, JALAL R, et al. Spinal needle navigation by tracked ultrasound snapshots[J]. IEEE transactions on biomedical engineering, 2016, 59(10): 2766–2772.
[15] 许奇, 王华彬, 周健, 等. 用于目标跟踪的智能群体优化滤波算法[J]. 智能系统学报, 2019, 14(4): 697–707
[16] 王伟, 胡子英, 岳佳男. 高效的MIMO雷达运动目标三维成像方法[J]. 通信学报, 2019, 40(7): 38–47
[17] OLIVEIRA T, AGUIAR A P, ENCARNACAO P. Moving path following for unmanned aerial vehicles with applications to single and multiple target tracking problems[J]. IEEE transactions on robotics, 2016, 32(5): 1062–1078.
[18] ZENG Y L, LAN J H, ZOU J L. Research on infrared dim-small target tracking algorithm with template sliding confidence constraint[J]. Acta armamentarii, 2017, 38(9): 1771–1778.
[19] 苏贇, 王挺, 姚辰, 等. 基于合作目标的无人机目标跟踪方法[J]. 机器人, 2019, 41(4): 425–432
[20] 李东方, 邓宏彬, 潘振华, 等. 基于改进蛇形曲线的蛇形机器人在流场中避障的轨迹跟踪控制律[J]. 机器人, 2019, 41(4): 433–442
[21] NIEDFELDT P C, BEARD R W. Convergence and complexity analysis of recursive-RANSAC: a new multiple target tracking algorithm[J]. IEEE transactions on automatic control, 2016, 61(2): 456–461.

备注/Memo

备注/Memo:
收稿日期:2020-05-13。
基金项目:教育部科技发展中心基金项目(2018A01002);中国博士后基金项目(2017M610852);吉林省社科基金重点项目(2016A5)
作者简介:李晓峰,男,教授,博士
通讯作者:李晓峰,E-mail:lixiaofeng@hiu.net.cn
更新日期/Last Update: 2020-11-27