Addressing wander effect in vehicle weight monitoring: An advanced hybrid weigh-in-motion system integrating computer vision and in-pavement sensors

Abstract

The paper addressed the challenges of the two main weigh-in-motion (WIM) systems through a hybrid WIM system, which combines embedded in-pavement sensor systems and computer vision to enhance accuracy while minimizing human intervention. To establish a cost-effective system, this study employed a standard camera integrated with computer vision technique to improve the precision of identifying wheel location and automatically adjust distance calibration, particularly in response to variations in camera angle. In addition, the system achieved real-time vehicle weight detection efficiently by addressing the vehicle wander effect concerns of the in-pavement sensors through camera using computer vision. The system can achieve weight accuracy notably exceeding ASTM 1318E-09 standards, with over 90% accuracy for distances below 0.1 m, demonstrating its potential value in assisting traffic analysis and pavement maintenance by offering real-time, accurate WIM measurements at a low cost for instrumentation.