A visual ergonomic assessment approach using Kinect and OWAS in real workplace environments

Ergonomics plays an important role and has contributed to sustainable development in many areas such as product design, architecture, health, safety, and workplace design. An ergonomic assessment is a crucial task in real workplace environments to prevent potential musculoskeletal disorders. Recently, visual ergonomic assessment has been widely utilized for skeleton analysis of human joints for body posture identification to deal with musculoskeletal disorders risks. However, posture identification has limitations in self-occlusion joint postures. This study presents a visual ergonomic assessment approach for posture identification in free- and self-occlusion conditions. For self-occlusion detection, as the main focus of this study, an algorithm is proposed to overcome this limitation to detect the joints' location when other relative joints block the joints. After the self-occlusion is detected, the location of a blocked joint is identified using the primary data collected in body data extraction in the joint location estimation process. Then, the identified location of the joint is used for posture identification in the free and self-occlusion detection process. The posture identification is based on the OWAS standard for posture and category identification. Finally, experimental results and performance evaluation are presented in individual and integrated procedures. In individual evaluation, the performance of the algorithm is reported for the self- and free-occlusion detection, posture, and category identification processes separately. The results are collected for the overall proposed approach in integrated evaluation, and the performance is measured using standard evaluation metrics. As experimental results show, the proposed approach can effectively detect the postures and identify the associated category in the OWAS standard for both self- and free occlusions.