Accuracy of a markerless motion capture system for postural ergonomic risk assessment in occupational practice

Established methods for postural ergonomic risk assessment in occupational practice are mostly time-consuming and need to be conducted by experts. Use of technology could improve postural ergonomic risk assessments with regard to time efficiency and accuracy. A study was conducted to assess the accuracy of a markerless motion capture system (Microsoft Kinect V2) compared to a marker-based motion capture system (Vicon Bonita). Angles of different body segments were analysed. The results show major inaccuracies of the markerless motion capture system for capturing axial trunk rotation (mean angular deviation of 14.04 degrees) indicating that potential health risks could be underestimated. Combined working postures of axial trunk rotation and arm anteversion show issues with self-occlusion. Based on the findings, it is discussed whether the detected inaccuracies for axial trunk rotation are likely to lead to overestimation or underestimation of potential health risks when conducting an ergonomic risk assessment.