基于Azure Kinect骨骼追踪的腕关节活动度测量方法

doi:10.3969/j.issn.1006-9771.2022.08.014

《中国康复理论与实践》 ›› 2022, Vol. 28 ›› Issue (8): 981-988.doi: 10.3969/j.issn.1006-9771.2022.08.014

基于Azure Kinect骨骼追踪的腕关节活动度测量方法

张文波¹,瞿畅¹(),周建萍²,张小萍¹,张啸天¹

1.南通大学机械工程学院,江苏南通市 226019
2.南通大学附属医院,江苏南通市 226001

收稿日期:2021-12-02 修回日期:2022-05-25 出版日期:2022-08-25 发布日期:2022-08-30
通讯作者: 瞿畅 E-mail:xu.ch@ntu.edu.cn
作者简介:张文波(1996-),男,汉族,河南开封市人,硕士研究生,主要研究方向：虚拟现实技术及应用。|瞿畅(1967-),女,汉族,江苏南通市人,硕士,教授,硕士研究生导师,主要研究方向：虚拟现实技术及应用。
基金资助:
南通市社会民生科技计划项目(MS12021022)

Wrist joint range of motion measurement based on Azure Kinect bone tracking

ZHANG Wenbo¹,QU Chang¹(),ZHOU Jianping²,ZHANG Xiaoping¹,ZHANG Xiaotian¹

1. School of Mechanical Engineering, Nantong University, Nantong, Jiangsu 226019, China
2. Affiliate Hospital of Nantong University, Nantong, Jiangsu 226001, China

Received:2021-12-02 Revised:2022-05-25 Published:2022-08-25 Online:2022-08-30
Contact: QU Chang E-mail:xu.ch@ntu.edu.cn
Supported by:
Nantong Social and Livelihood Science and Technology Plan(MS12021022)

摘要/Abstract

摘要：

目的使用体感捕捉设备Azure Kinect,实现腕关节活动度的自动测量。

方法利用Azure Kinect识别人体肘、腕、手掌、指尖等关节点的空间坐标,通过正交试验确定最佳测量位姿;运用霍尔特双参数指数平滑法、骨长约束法对获取的关节点数据进行平滑处理,消除抖动,通过空间向量关系计算多帧角度的平均值,实现腕关节活动度的自动测量。2021年5月至10月,采用上述方法对5例健康受试者的腕关节各科目进行5次×10组测量试验。

结果得到腕关节各科目的 $R 1$ 值均处于83和127之间,且数值居中。该方法测量结果与增强现实技术尺子、二维影像测量结果有极强相关性(r > 0.990),最大误差在1.61°以内。

结论 Azure Kinect腕关节角度测量方式与传统测量方式测量结果无明显差异,可以应用于腕关节活动度的测量,并满足实时测量要求。

关键词: 腕, 关节活动度, 霍尔特双参数指数平滑法, 骨长约束, 骨骼追踪, Azure Kinect

Abstract:

Objective To automatically measure the wrist range of motion using the somatosensory capture device Azure Kinect.

Methods Azure Kinect was used to recognize the spatial coordinates of human elbow, wrist, palm, fingertip and other joint points, and the best measurement posture was determined by orthogonal experiment. Holt's double-parameter exponential smoothing method and bone length constraint method were used to smooth the joint point data to eliminate the jitter. The average angle of multiple frames was calculated through the spatial vector relationship to realize the automatic measurement of wrist joint range of motion. From May to October, 2021, 5 times × 10 groups of measurement were performed on each subject of wrist joint of five healthy subjects using the above method.

Results The R1 of each subject of wrist joint was between 83 and 127, and the value was in the middle. This method contrasted well with measurements made on augmented reality ruler, 2D images (r > 0.990), and the maximum error was within 1.61°.

Conclusion The wrist joint angle measurement modality using Azure Kinect is consistent to conventional measurement modality, and can meet the requirements of real-time measurement.

Key words: wrist, range of motion, Holt's double-parameter exponential smoothing method, bone length constraint, bone tracking, Azure Kinect

中图分类号:

R496

张文波,瞿畅,周建萍,张小萍,张啸天. 基于Azure Kinect骨骼追踪的腕关节活动度测量方法[J]. 《中国康复理论与实践》, 2022, 28(8): 981-988.

ZHANG Wenbo,QU Chang,ZHOU Jianping,ZHANG Xiaoping,ZHANG Xiaotian. Wrist joint range of motion measurement based on Azure Kinect bone tracking[J]. 《Chinese Journal of Rehabilitation Theory and Practice》, 2022, 28(8): 981-988.

图/表 12

图1

图2

表1

因素水平表"

水平	距离(A)/mm	角度(B)/°	姿势(C)
1	1000	$-$ 45°	a
2	2500	0°	b
3	4000	45°	c

表1

表2

表3

试验结果分析"

项目	A		B	C
$K 1$	65.37	122.51	50.48	71.40
$K 2$	91.44	143.18	138.45	136.87
$K 3$	179.78	70.90	147.66	128.32
R	114.41	72.28	97.18	65.47

表3

图3

图4

图5

图6

表4

表5

图7

参考文献 19

[1]	KIM J B, KIM J C, HWANG D J. A study on the robot structure of hand for the rehabilitation training of stroke patients[J]. J Biomed Eng Res, 2019, 40(3): 116-124.
[2]	WAGNER E R, CONTI MICA M, SHIN A Y. Smartphone photography utilized to measure wrist range of motion[J]. J Hand Surg Eur Vol, 2018, 43(2): 187-192. doi: 10.1177/1753193417729140
[3]	WIRTH M A, FISCHER G, REISSNER L, et al. Comparison of a new inertial sensor based system with an optoelectronic motion capture system for motion analysis of healthy human wrist joints[J]. Sensors (Basel), 2019, 19(23): 5297. doi: 10.3390/s19235297
[4]	HUANG B, LI M Y, MEI T, et al. Wearable stretch sensors for motion measurement of the wrist joint based on dielectric elastomers[J]. Sensors (Basel), 2017, 17(12): 2708. doi: 10.3390/s17122708
[5]	YANG X C, YAN J P, FANG Y F, et al. Simultaneous prediction of wrist/hand motion via wearable ultrasound sensing[J]. IEEE Trans Neural Syst Rehabil Eng, 2020, 28(4): 970-977. doi: 10.1109/TNSRE.2020.2977908
[6]	李培丰, 林愉添, 朱君毅, 等. 一种新的智能手机应用程序AR尺子测量腕关节活动度的可靠性和同时效度评估[J]. 中华手外科杂志, 2020, 36(6): 425-429.
	LI P F, LIN Y T, ZHU J Y, et al. Reliability and concurrent validity of a new smartphone application AR ruler for measuring active wrist range of motion[J]. Chin J Hand Surg, 2020, 36(6): 425-429.
[7]	KIM T S, PARK D D, LEE Y B, et al. A study on the measurement of wrist motion range using the iPhone 4 gyroscope application[J]. Ann Plast Surg, 2014, 73(2): 215-218. doi: 10.1097/SAP.0b013e31826eabfe
[8]	POURAHMADI M R, EBRAHIMI TAKAMJANI I, SARRAFZADEH J, et al. Reliability and concurrent validity of a new iPhone goniometric application for measuring active wrist range of motion: a cross‐sectional study in asymptomatic subjects[J]. J Anat, 2017, 230(3): 484-495. doi: 10.1111/joa.12568
[9]	MCHUGH B P, MORTON A M, AKHBARI B, et al. Accuracy of an electrogoniometer relative to optical motion tracking for quantifying wrist range of motion[J]. J Med Eng Technol, 2020, 44(2): 49-54. doi: 10.1080/03091902.2020.1713240
[10]	KIM K G, PARK C S, JEON S H, et al. Feasibility of a new desktop motion analysis system with a video game console for assessing various three-dimensional wrist motions[J]. Clin Orthop Surg, 2018, 10(4): 468-478. doi: 10.4055/cios.2018.10.4.468
[11]	瞿畅, 丁晨, 王君泽, 等. 基于Kinect体感交互技术的上肢关节活动度测量方法[J]. 中国生物医学工程学报, 2014, 33(1): 16-21.
	QU C, DING C, WANG J Z, et al. A method to measure the range of motion of human upper limbs based on Kinect somatosensory interaction technology[J]. Chin J Biomed Eng, 2014, 33(1): 16-21.
[12]	王伟伟, 郭远其, 高卉, 等. 基于深度图像的关节活动度测量及其结果评价[J]. 中国医学物理学杂志, 2016, 33(3): 262-269.
	WANG W W, GUO Y Q, GAO H, et al. Measuring joint range of motion and evaluating measured results based on depth images[J]. Chin J Med Phys, 2016, 33(3): 262-269.
[13]	ALBERT J A, OWOLABI V, GEBEL A, et al. Evaluation of the pose tracking performance of the Azure Kinect and Kinect v2 for gait analysis in comparison with a gold standard: a pilot study[J]. Sensors (Basel), 2020, 20(18): 5104. doi: 10.3390/s20185104
[14]	LEE C, KIM J, CHO S, et al. Development of real-time hand gesture recognition for tabletop holographic display interaction using Azure Kinect[J]. Sensors (Basel), 2020, 20(16): 4566. doi: 10.3390/s20164566
[15]	张弛. Kinect关节数据处理算法及在机器人体感控制中的应用[D]. 武汉: 华中科技大学, 2017.
	ZHANG C. A Kinect joint data processing algorithm and its application in robot motion-sensing control[D]. Wuhan: Huazhong University of Science and Technology, 2017.
[16]	蒋成林. 霍尔特指数平滑法参数的优选[J]. 统计教育, 2004, 12(4): 13-15.
	JIANG C L. The optimization of parameters about holt exponential smoothing[J]. Stat Educ, 2004, 12(4): 13-15.
[17]	GOPURA R, BANDARA D, KIGUCHI K, et al. Developments in hardware systems of active upper-limb exoskeleton robots: a review[J]. Rob Auton Syst, 2016, 75 (Pt.B): 203-220.
[18]	ZHU Y, LU W, GAN W H, et al. A contactless method to measure real-time finger motion using depth-based pose estimation[J]. Comput Biol Med, 2021, 131(12): 104282. doi: 10.1016/j.compbiomed.2021.104282
[19]	瞿畅, 沈芳, 于陈陈, 等. 基于Kinect深度图像的腕部及手指活动度测量方法[J]. 中国生物医学工程学报, 2016, 35(5): 626-630.
	QU C, SHEN F, YU C C, et al. Measurement for detecting the range of motion of fingers and wrist based on depth image from Kinect[J]. Chin J Biomed Eng, 2016, 35(5): 626-630.

试验号	A		B	C	抖动程度
1	1	1	1	1	4.80
2	1	2	2	2	35.35
3	1	3	3	3	25.22
4	2	1	2	3	34.30
5	2	2	3	1	39.03
6	2	3	1	2	18.11
7	3	1	3	2	83.41
8	3	2	2	3	68.80
9	3	3	1	1	27.57

科目	掌屈		背伸		尺偏		桡偏
科目	A	B	A	B	A	B	A	B
男1左	101	105	103	105	104	110	102	104
男2左	109	106	107	108	108	103	106	107
男3右	102	103	101	101	109	109	102	104
女1左	101	104	105	106	103	107	106	103
女2右	104	103	105	103	103	105	95	102

项目	掌屈	背伸	尺偏	桡偏
r_a	0.998	0.996	0.996	0.990
r_b	0.998	0.997	0.997	0.992

基于Azure Kinect骨骼追踪的腕关节活动度测量方法

Wrist joint range of motion measurement based on Azure Kinect bone tracking

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 12

参考文献 19

相关文章 15

Metrics

本文评价

推荐阅读 0

[1]	徐翰林, 胡国炯, 郑绍城, 曾晓文, 曾宪华, 邵文启. 脑卒中偏瘫患者膝过伸步态与下肢肌肉激活功能的相关性[J]. 《中国康复理论与实践》, 2023, 29(2): 151-155.
[2]	李思佳,张琦,何艳,刘前进,张燕庆,周天添. 痉挛型双瘫脑瘫儿童粗大运动功能与脊柱活动度的相关性[J]. 《中国康复理论与实践》, 2022, 28(10): 1211-1216.
[3]	王冲,杜良杰,王志功,李建军,李军,刘宏炜,丛芳,林歆. 微电子肌电桥技术对颈脊髓损伤患者桡侧腕长伸肌运动功能康复的效果[J]. 《中国康复理论与实践》, 2021, 27(6): 698-705.
[4]	马善新,许建文,马楠,苏义,林伟山,陶广林. 佩戴腕托对脑卒中后偏瘫手功能康复的影响[J]. 《中国康复理论与实践》, 2020, 26(3): 306-309.
[5]	黄楚红,曾庆,陈欢,李宇淇,黄国志. 复合式腕手矫形器辅助疗法对脑卒中后腕手功能障碍的效果[J]. 《中国康复理论与实践》, 2020, 26(12): 1446-1452.
[6]	杨绯, 潘钰, 吴琼, 徐泉, 李欣, 张丽春. 踝扭伤后关节生物力学和本体感觉变化特征及相关性[J]. 《中国康复理论与实践》, 2019, 25(12): 1365-1369.
[7]	杨名珍, 黄崧华, 白玉龙. 动态腕手矫形器应用于慢性期脑卒中患者上肢及手功能康复的疗效[J]. 《中国康复理论与实践》, 2019, 25(11): 1361-1364.
[8]	刘静娅, 刘璇, 黄富表, 杨絮. 作业治疗在腕管综合征中的应用[J]. 《中国康复理论与实践》, 2017, 23(5): 563-566.
[9]	惠寒青;张玲莉;喻大力;王荣;余竹生. 新型颈椎三维电子关节角度仪的效度及重复测试信度[J]. 《中国康复理论与实践》, 2015, 21(11): 1329-1332.
[10]	刘洪举;丘云锋;董玲;陈飞;覃婷;丁一航. 悬吊训练在膝关节置换术后康复中的作用①[J]. 《中国康复理论与实践》, 2015, 21(08): 947-952.
[11]	叶大勇;张希彬;李宝. 改良腕手关节矫形器对脑卒中偏瘫患者腕手关节屈肌痉挛及运动功能的疗效[J]. 《中国康复理论与实践》, 2015, 21(07): 811-815.
[12]	李洋;朱向阳;黄怀宇. 腕管综合征的诊断进展[J]. 《中国康复理论与实践》, 2013, 19(3): 246-249.
[13]	赵雅宁;马素慧;李建民;郭霞;刘丹;郝正伟. 步态训练机器人对脑卒中偏瘫患者步行功能的影响[J]. 《中国康复理论与实践》, 2013, 19(1): 57-59.
[14]	刘振寰;张丽红;赵勇. 五行体感音乐对痉挛型脑性瘫痪患儿肌张力的影响[J]. 《中国康复理论与实践》, 2013, 19(08): 771-774.
[15]	郑志新;高谦;王军;侯京山;姜岳波;马玲;陈国辉. 气动弹道式体外冲击波治疗慢性非特异性下腰痛的效果[J]. 《中国康复理论与实践》, 2013, 19(07): 666-668.