This study aims to develop an image processing model for assessing truck drivers’ sitting posture based on ergonomic principles using the Rapid Upper Limb Assessment (RULA) method, employing Computer Vision technology integrated with MediaPipe Pose Estimation. The reliability of the model was evaluated by comparing joint angle measurements with those obtained by an expert assessor. The developed model is capable of detecting six body landmarks from lateral photographs and computing four joint angles, namely the neck, trunk, upper arm, and lower arm. A Backward Offset technique was developed to adjust landmark positions to correspond with anatomical reference points as specified by the RULA criteria. Testing with a sample of 22 participants revealed that the Intraclass Correlation Coefficients (ICC) for joint angles ranged from 0.777 to 0.995 (good to excellent), and the ICC values for RULA scores ranged from 0.625 to 1.000 (moderate to excellent). These findings demonstrate that the model possesses high reliability in estimating joint angles and can serve as a preliminary ergonomic risk screening tool for truck drivers.

Keywords

Posture Assessment, Computer Vision, Truck Drivers, Ergonomics, Musculoskeletal Disorders.