AUTOMATIC VIDEO-BASED MONITORING OF TASK EXECUTION IN PHYSICALLY DEMANDING JOBS TO PREVENT MUSCULOSKELETAL PATHOLOGIES

This thesis focuses on the prevention of Work-Related Musculoskeletal Disorders (WMSDs) through the analysis of worker movements using motion capture systems. By examining key joint angles—elbow, knee, and hip—across correct and incorrect task executions, unsupervised learning techniques like k-means clustering and Otsu’s thresholding were applied to classify and identify risky postures. The results highlight significant differences between correct and incorrect movements, offering insights into potential injury prevention strategies.