Non-contact weigh-in-motion approach with an improved multi-region of interest method

Abstract

Vehicle weight monitoring is increasingly essential for traffic safety and infrastructure maintenance. Compared with traditional weigh-in-motion techniques, the non-contact weigh-in-motion (NC-WIM) has the advantages of non-intrusive installation, low cost, and rich information conveyance. However, most existing procedures for NC-WIM tend to be computationally intensive and time-consuming. To improve the data processing efficiency in NC-WIM, a multi-region of interest (ROI) method is proposed. With this method, the latency for tire deformation extraction can be decreased from more than 1 min to less than 0.15 s. Meanwhile, the multi-ROI method is robust to daytime lighting variations and works well under both backlighting and frontlighting on cloudy and sunny days. Based on the outcomes of this study, the ability of NC-WIM systems to cope with heavy traffic and complex illumination can be significantly enhanced. The improved NC-WIM is promising to provide long-term data for studies such as structural health monitoring, pavement design, and low-carbon transportation.