基于双Kinect传感器的三维人体姿态跟踪方法

doi:10.16182/j.issn1004731x.joss.19-0017

系统仿真学报 ›› 2020, Vol. 32 ›› Issue (8): 1446-1454.doi: 10.16182/j.issn1004731x.joss.19-0017

基于双Kinect传感器的三维人体姿态跟踪方法

李琪^1,2, 王向东³, 李华^1,2

1.中国科学院计算技术研究所院智能信息处理重点实验室,北京 100190;
2.中国科学院大学,北京 100049;
3.国家体育总局体育科学研究所,北京 100061

收稿日期:2019-01-09 修回日期:2019-03-12 出版日期:2020-08-18 发布日期:2020-08-13
作者简介:李琪(1992-),男,安徽蚌埠,博士生,研究方向为计算机图形学与人体动作识别。
基金资助:
国家重点研发计划(2017YFB1002703),国家重点基础研究规划(2015CB554507),国家自然科学基金(61379082),国家体育总局体育科学研究所合作项目(基本17-01,基本17-28,基本18-19)

3D Human Pose Tracking Approach Based on Double Kinect Sensors

Li Qi^1,2, Wang Xiangdong³, Li Hua^1,2

1. Key Laboratory of Intelligent Information Processing, Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100190, China;
2. University of Chinese Academy of Sciences, Beijing 100049, China;
3. China Institute of Sport Science, General Administration of Sport, Beijing 100061, China

Received:2019-01-09 Revised:2019-03-12 Online:2020-08-18 Published:2020-08-13

摘要/Abstract

摘要： 三维人体姿态估计与跟踪是机器视觉中的热点问题。为了提高人体姿态跟踪的稳定性,以人体关节点自由度向量的方法表示人体运动姿态,使用无迹卡尔曼滤波方法对人体姿态进行跟踪,并搭建了一套基于双Kinect传感器的人体姿态跟踪系统。对比于传统的人体运动捕捉系统,该系统可实现在简单动作下准确稳定的三维人体姿态跟踪,在复杂动作下反映运动过程的特殊性质,并用于运动生物力学的性能评估。

关键词: 三维人体姿态跟踪, Kinect传感器, 关节点自由度, 无迹卡尔曼滤波, 运动生物力学

Abstract: 3D human pose estimation and tracking is a popular problem of machine vision. In order to improve the stability of human pose tracking, the human body motion posture is represented by the human joint DOF (Degree of Freedom) vector, the human pose is tracked by the unscented Kalman filter method, and a human pose tracking system based on double Kinect sensors is built. Compared with the traditional human motion capture system, the system can accurately and stably track 3D human pose under simple movements, reflect the special nature of the motion process under complex movements, and can be used in the performance evaluation of sports biomechanics.

Key words: 3D human pose tracking, Kinect sensor, human joint DOF, unscented Kalman filter, sports biomechanics

中图分类号:

TP391

李琪, 王向东, 李华. 基于双Kinect传感器的三维人体姿态跟踪方法[J]. 系统仿真学报, 2020, 32(8): 1446-1454.

Li Qi, Wang Xiangdong, Li Hua. 3D Human Pose Tracking Approach Based on Double Kinect Sensors[J]. Journal of System Simulation, 2020, 32(8): 1446-1454.

参考文献 13

[1]	Moeslund T B, Hilton A, Krüger V.A Survey of Advances in Vision-based Human Motion Capture and Analysis[J]. Computer Vision and Image Understanding (S1077-3142), 2006, 104(2/3): 90-126.
[2]	Shotton J, Girshick R, Fitzgibbon A, et al.Efficient Human Pose Estimation from Single Depth Images[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence (S0162-8828), 2013, 35(12): 2821-2840.
[3]	Wei X, Zhang P, Chai J.Accurate Real-time Full-body Motion Capture Using a Single Depth Camera[J]. ACM Transactions on Graphics (S0730-0301), 2012, 31(6): 188.
[4]	Shuai L, Li C, Guo X, et al.Motion Capture with Ellipsoidal Skeleton Using Multiple Depth Cameras[J]. IEEE Transactions on Visualization and Computer Graphics (S1077-2626), 2017, 23(2): 1085-1098.
[5]	Guess T M, Razu S, Jahandar A, et al.Comparison of 3D Joint Angles Measured with the Kinect 2.0 Skeletal Tracker versus a Marker-based Motion Capture System[J]. Journal of Applied Biomechanics (S1065-8483), 2017, 33(2): 176-181.
[6]	Moon S, Park Y, Ko D W, et al.Multiple Kinect Sensor Fusion for Human Skeleton Tracking Using Kalman Filtering[J]. International Journal of Advanced Robotic Systems (S1729-8814), 2016, 13(2): 65.
[7]	Joo H, Simon T, Li X, et al.Panoptic Studio: A Massively Multi-view System for Social Interaction Capture[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence (S0162-8828), 2019, 41(1): 190-204.
[8]	Ionescu C, Papava D, Olaru V, et al.Human3. 6m: Large Scale Datasets and Predictive Methods for 3D Human Sensing in Natural Environments[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence (S0162-8828), 2014, 36(7): 1325-1339.
[9]	Zhou X, Huang Q, Sun X, et al.Towards 3D Human Pose Estimation in the Wild: A Weakly-supervised Approach[C]//IEEE International Conference on Computer Vision. New York: IEEE Press, 2017: 398-407.
[10]	Bregler C, Malik J.Tracking People with Twists and Exponential Maps[C]//IEEE Conference on Computer Vision and Pattern Recognition. New York: IEEE Press, 1998: 8-15.
[11]	Taylor G W, Hinton G E, Roweis S T.Modeling Human Motion Using Binary Latent Variables[C]//Advances in Neural Information Processing Systems. Cambridge Massachusetts: MIT Press, 2007: 1345-1352.
[12]	Wan E A, van Der Merwe R. The Unscented Kalman Filter for Nonlinear Estimation[C]//Adaptive Systems for Signal Processing, Communications, and Control Symposium 2000. New York: IEEE Press, 2000: 153-158.
[13]	van Der Merwe R. Sigma-point Kalman Filters for Probabilistic Inference in Dynamic State-space Models [D]. 2004.

基于双Kinect传感器的三维人体姿态跟踪方法

3D Human Pose Tracking Approach Based on Double Kinect Sensors

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献 13

相关文章 5

编辑推荐

Metrics

本文评价

[1]	翟志强, 许进亮, 袁皓, 周永明, 赵龙飞, 王世谦. 自适应巡航系统车辆跟驰控制策略仿真[J]. 系统仿真学报, 2020, 32(5): 885-891.
[2]	贾宇涵, 吴建平, 许明, 刘明宇, 何磊. 基于FlowSIM的动态交通分配系统离线参数标定研究[J]. 系统仿真学报, 2017, 29(9): 2198-2205.
[3]	周彦, 胡岚, 王冬丽. 基于IMM-UKF-CS的混合环境目标跟踪[J]. 系统仿真学报, 2017, 29(1): 43-48.
[4]	焦俊, 孔文, 辜丽川, 王强, 高雅, 袁晨晨. 基于UKF和SMO农用履带机器人滑动参数计算[J]. 系统仿真学报, 2015, 27(7): 1577-1583.
[5]	毛雁明, 章立亮. 基于Kinect深度信息的手势分割与识别[J]. 系统仿真学报, 2015, 27(4): 830-835.