[1]王子翱,王卫星,陈靖翕,等.基于MMPose的三维作业姿势风险与角度可靠性评估*[J].中国安全生产科学技术,2023,19(12):200-206.[doi:10.11731/j.issn.1673-193x.2023.12.027]
WANG Ziao,WANG Weixing,CHEN Jingxi,et al.Risk and angle reliability assessment of 3D operational postures based on MMPose[J].JOURNAL OF SAFETY SCIENCE AND TECHNOLOGY,2023,19(12):200-206.[doi:10.11731/j.issn.1673-193x.2023.12.027]
点击复制
基于MMPose的三维作业姿势风险与角度可靠性评估*
WANG Ziao,WANG Weixing,CHEN Jingxi,et al.Risk and angle reliability assessment of 3D operational postures based on MMPose[J].JOURNAL OF SAFETY SCIENCE AND TECHNOLOGY,2023,19(12):200-206.[doi:10.11731/j.issn.1673-193x.2023.12.027]
《中国安全生产科学技术》[ISSN:1673-193X/CN:11-5335/TB]
- 卷:
- 19
- 期数:
- 2023年12期
- 页码:
- 200-206
- 栏目:
- 职业安全卫生管理与技术
- 出版日期:
- 2023-12-31
文章信息/Info
- Title:
- Risk and angle reliability assessment of 3D operational postures based on MMPose
- 文章编号:
- 1673-193X(2023)-12-0200-07
- 关键词:
- 肌肉骨骼疾患（WMSDs）; 三维人体姿态识别; 人体工学姿势风险评估
- Keywords:
- work-related musculoskeletal disorders (WMSDs); three-dimensional human posture recognition; ergonomic posture risk assessment
- 分类号:
- X948；X912.9
- 文献标志码:
- A
- 摘要:
- 为避免作业人员因工作姿势不准确或过度负荷而患上工作相关的肌肉骨骼疾患（WMSDs），提出1种非侵入式的三维姿势风险评估方法。首先，构建三维人体姿态估计模型；其次，使用三维姿态识别技术提取人体的三维坐标并计算出所对应的关节角度；最后，将关节角度作为输入放入REBA，RULA评分库中进行作业姿势风险的自动评估，引入身体角度可靠性评估，可有效过滤一些采集质量较低的人体姿势视角数据，提高作业姿势估计的准确性。研究结果表明：三维姿势风险评估的识别精度优于Kinect，对比关节识别精度均方根误差(RMSE)更小；且方法具有可行性，与3位专家比较类内相关系数（ICC值）均大于0.846；并且通过身体可靠性评估，REBA，RULA的分数评估精确度都提升至94%。研究结果可为减低作业人员职业健康风险，提高工作效率提供参考与借鉴。
- Abstract:
- In order to avoid the work-related musculoskeletal disorders (WMSDs) among operators due to inaccurate work postures or overloading,a non-invasive 3D posture risk assessment method was proposed.Firstly,a 3D human posture estimation model was constructed.Secondly,the 3D coordinates of the human body were extracted and the corresponding joint angles were calculated using 3D posture recognition technology.Finally,the joint angles were put into REBA and RULA scoring libraries as inputs for the automatic assessment of operational posture risk.The introduction of body angle reliability assessment could effectively filter some human posture viewpoint data with lower acquisition quality,and improve the accuracy of operational posture estimation.The results show that the recognition accuracy of 3D posture risk assessment is better than Kinect,and the root mean square error (RMSE) of the recognition accuracy of the comparison joints is smaller; and the method is feasible.The intraclass correlation coefficients (ICC values) are all greater than 0.846 when compared with three experts,and the accuracy of the score estimation of REBA and RULA are both improved to 94% by the body reliability assessment.The research results can provide reference for reducing the occupational health risk of workers and improving work efficiency.
参考文献/References:
［1］陈青松.工作相关肌肉骨骼疾患及其防控［J］.环境与职业医学，2023，40（1）:1-5
CHEN Qingsong.Work-related musculoskeletal disorders and their prevention and control［J］.Journal of Environmental and Occupational Medicine,2023,40(1):1-5.
［2］张瑞秋，李泽，李育奇.肌肉骨骼疾患风险评估方法趋势研究［J］.包装工程，2020，41（14）:49-60.
ZHANG Ruiqiu，LI Ze，LI Yuqi.Trends of risk assessment methods for musculoskeletal diseases［J］.Packaging Engineering,2020,41(14):49-60.
［3］陈星,常原,郑颖琇,等.基于AlphaPose与REBA的手工搬运作业姿势风险评估方法［J］.中国安全生产科学技术,2022,18(7):232-238.
CHEN Xing,CHANG Yuan,ZHENG Yingxiu,et al.Risk assessment method on work posture of manual handling based on AlphaPose and REBA［J］.Journal of Safety Science and Technology,2022,18(7):232-238.
［4］郝培，李医博，武姗姗,等.机械制造行业员工工作有关肌肉骨骼疾患影响因素分析［J］.中华劳动卫生职业病杂志，2020，38（3）:187-192.
HAO Pei，LI Yibo，WU Shanshan,et al.Investigation and analysis of work-related occupational musculoskeletal disorders and associated risk factors of manufacturing workers［J］.Chinese Journal of Industrial Hygiene and Occupational Diseases,2020,38(3):187-192.
［5］曹磊，王忠旭，贾宁,等.汽车制造工人肌肉骨骼疾患及不良工效学因素的调查与分析［J］.中国工业医学杂志，2020，33（3）:206-210.
CAO Lei，WANG Zhongxu，JIA Ning,et al.Investigation and analysis of musculoskeletal disorders and adverse ergonomic factors automobile workers［J］.Chinese Journal of Industrial Medicine,2020,33(3):206-210.
［6］LIN S C,LIN L L,LIU C J,et al.Exploring the factors affecting musculoskeletal disorders risk among hospital nurses［J］.PLoS One，2020，15（4）:e231319.
［7］YONG X T,LI F Y,GE H，et al.A cross-sectional epidemiological survey of work-related musculoskeletal disorders and analysis of its influencing factors among coal mine workers in Xinjiang［J］.BioMed Research International，2020，2020:3164056.
［8］YUAN H,ZHOU Y.Ergonomic assessment based on monocular RGB camera in elderly care by a new multi-person 3D pose estimation technique (ROMP)［J］.International Journal of Industrial Ergonomics，2023，95:103440.
［9］MCATAMNEY L,CORLETT E N.RULA:a survey method for the investigation of work-related upper limb disorders［J］.Applied Ergonomics，1993，24（2）:91-99.
［10］HIGNETT S,MCATAMNEY L.Rapid entire body assessment (REBA)［J］.Applied Ergonomics，2000，31（2）:201-205.
［11］MARIA L N,DUARTE F,CARLOS F,et al.Posture risk assessment in an automotive assembly line using inertial sensors［J］.IEEE Access，2022，10:83221-83235.
［12］WEI FANG,MINGYU FU,LIANYU ZHENG.Continuous ergonomic risk perception for manual assembly operations using wearable multi-sensor posture estimation［J］.Assembly Automation，2022，42（2）:209-217.
［13］HAROLD V J,OREJUELA I Z,BAYDAL-BERTOMEU J M.Study of the ergonomic risk in operators of an assembly line using the RULA method in real working conditions through the application of a commercial sensor［J］.Materials Today:Proceedings，2022，49:122-128.
［14］SUN K,XIAO B,LIU D,et al.Deep high-resolution rep-resentation learning for human pose estima-tion［C］//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition,2019:5693-5703.
［15］蔡哲栋，应娜，郭春生,等.YOLOv3剪枝模型的多人姿态估计［J］.中国图象图形学报，2021,26(4):837-846.
CAI Zhedong,YING Na,GUO Chunsheng,et al.Research on multiperson pose estimation combined with YOLOv3 pruning model［J］.Chinese Journal of Image Graphics,2021,26(4):837-846.
［16］ZHAO T T.Research on miners’ human posture detection algorithm based on MMPose［J］.Academic Journal of Computing & Information Science，2023，6（4）:57-62.
［17］GIRSHICK R.Fast r-cnn［C］//Proceedings of the IEEE International Conference on Computer Vision,2015:1440-1448.
［18］CHARLES J,PFISTER T,MAGEE D,et al.Personalizing human video pose estimation［C］//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,2016:3063-3072.
［19］LIN T Y,MAIRE M,BELONGIE S,et al.Microsoft coco:Common objects in context［C］//Computer Vision-ECCV 2014:13th European Conference,Zurich,Switzerland,2014,Proceedings,Part V 13.Springer International Publishing,2014:740-755.
相似文献/References:
备注/Memo
- 备注/Memo:
-
收稿日期： 2023-07-21
* 基金项目： 贵州省科技计划项目（黔科合基础-ZK〔2023〕重点015，黔科合基础〔2020〕1Y262)
作者简介： 王子翱，硕士研究生，主要研究方向为人机工程学，人体姿态估计。
通信作者： 王卫星，博士，副教授，主要研究方向为设计理论与方法，交互设计。
