Using computer-vision in digital human modelling tools to redesign workstations
2026 (English)Independent thesis Basic level (degree of Bachelor), 20 credits / 30 HE credits
Student thesis
Abstract [en]
Work-related musculoskeletal disorders (WMSDs) remain a significant issue in industrial workstations, negatively affecting workers’ well-being and productivity. This project focused on the redesign of a workstation in which harmful ergonomics conditions were identified, with the objective of reducing workers’ exposure to WMSDs through the integration of computer-vision (CV) and Digital Human Modelling (DHM) tools.
Motion capture (MoCap) data were obtained from video recordings using the RTMDet, FastSAM3D, and SAM3DBody pipelines. The extracted data were implemented in the DHM software IPS IMMA to evaluate workers’ poses through the ergonomics evaluation methods AFF, REBA, and QEC, allowing the identification of the most critical ergonomics risk factors. In parallel, CV-based techniques were also employed to generate 3D objects from images, facilitating the creation of a virtual representation of the workstation within the DHM environment.
Based on the ergonomics analysis, several redesign concepts were developed through brainstorming, concept generation, and concept evaluation methodologies. The selected redesign proposal was implemented and simulated in IPS IMMA, where a final ergonomics evaluation was conducted to validate the proposed improvements and evaluate their effectiveness in reducing exposure to harmful poses.
Place, publisher, year, edition, pages
2026. , p. x, 67
Keywords [en]
Ergonomics evaluation methods, MoCap, industry 5.0, human-centric, CV-based MoCap systems, IMUs-based MoCap systems, workstation redesign, WMSDs, REBA, AFF, QEC
National Category
Production Engineering, Human Work Science and Ergonomics
Identifiers
URN: urn:nbn:se:his:diva-26529OAI: oai:DiVA.org:his-26529DiVA, id: diva2:2073670
Subject / course
Product Design Engineering
Supervisors
Examiners
2026-06-162026-06-162026-06-16Bibliographically approved