Automatic real-time occupational posture evaluation and select corresponding ergonomic assessments.

Area of Science:

• Ergonomics and Occupational Health
• Biomechanical Analysis
• Computer Vision in Healthcare

Background:

• Current occupational risk assessments (e.g., REBA, RULA, OWAS) lack real-time feedback, delaying injury prevention.
• Musculoskeletal disorders are a significant concern in various work environments.
• Image-based motion capture offers potential for objective and immediate biomechanical analysis.

Purpose of the Study:

• To develop an automated system for selecting appropriate ergonomic assessment scales.
• To calculate musculoskeletal disorder risk scores using joint angle data from image-based motion capture.
• To enable real-time risk assessment for occupational injury prevention.

Main Methods:

• Utilized OpenPose for image-based motion capture to extract joint angle information.
• Developed a decision tree algorithm to automatically select assessment scales based on joint angles.
• Tested the system on 15 operation videos across six work types (maintenance, handling, assembly, cleaning, office work, driving).

Main Results:

• The decision tree accurately selected ergonomic assessment methods consistent with expert recommendations.
• High-risk postures were identified immediately, allowing for prompt intervention.
• The system demonstrated effectiveness across diverse occupational tasks.

Conclusions:

• The developed system provides rapid, real-time ergonomic assessment for musculoskeletal disorder risk.
• Automated scale selection and risk scoring enhance the applicability of ergonomic assessments in industrial settings.
• This approach facilitates proactive occupational injury prevention through immediate posture evaluation.