Research on Construction of Digital Twin System for Bridge Vehicle Load Identification

In order to improve the intelligent level of bridge operation and maintenance, breaking through the bottleneck of real-time positioning of moving vehicle load on bridge deck, This research combines image recognition, target tracking, data fusion and other technologies, a digital twin system for real-time identification of vehicle load position on bridge deck is developed. Firstly, the virtual bridge vehicle operation platform is developed, the multi-angle, multi-environment and multi-type vehicle data sets are obtained; Secondly, A bridge global image information fusion technology based on vision stitching is proposed, combined with the improved deep learning algorithm, the real-time recognition and tracking of vehicles in the whole bridge are realized; Furthermore, integrated bridge camera real-time image data and bridge dynamic weighing system, a digital twin system for real-time identification of vehicle load position on bridge deck is constructed; Finally, the above system is applied to Baijusi Yangtze River Bridge, realized the real-time identification of vehicles and vehicle load positions of Baijusi Yangtze River Bridge.