摘要
为有效预防工人肌肉骨骼疾病,提出基于深度神经网络（DNN）的作业姿势评估方法。利用人体关节点估计DNN模型,检测现场视频中工人的身体姿势;通过提取的骨骼关节点的空间位置,计算不同身体部位的姿势角度,并依据快速全身评估法（REBA）中的身体姿势角度与危害程度的关系,自动评估工人作业姿势的风险水平;以建筑工人典型施工作业姿势为例,验证该方法的有效性和实用性。结果表明：该方法能够克服身体遮挡、设备分辨率、光照条件等影响,准确检测关节点的位置;可自动连续评估工人的作业姿势,评估效率较高。
In order to effectively prevent the occurrence of occupational musculoskeletal disorder,an approach was developed for assessing the risk levels of working postures based on deep neural network. The spatial locations of joints were estimated from on-site video through DNN,then the REBA was employed based on the calculating limb angels for the ergonomic analysis. The method was verified by analyzing typical construction working postures. The results show that the developed methodology can automatically and continually analyze sequence of postures and satisfies the recording conditions like dim light,occlusion of the partial body and low resolution of camera,and that the developed approach improves the traditional REBA via DNN,leading to higher efficiency of assessment for improving professional health of workers.
作者
熊若鑫
宋元斌
王宇轩
XIONG Ruoxin;SONG Yuanbin;WANG Yuxuan(School of Naval Architecture,Ocean ＆ Civil Engineering,Shanghai Jiaotong University,Shanghai 200240,China;School of Transportation,Southeast University,Nanjing Jiangsu 211189,China)
出处
《中国安全科学学报》
CAS
CSCD
北大核心
2018年第5期105-110,共6页
China Safety Science Journal
基金
国家自然科学基金资助(71271137)
浦东新区卫计委重点学科群建设项目(PWZxq2017-16)
上海交通大学文理交叉基金资助(15JCMY09)
