ABSTRACT Timely and accurate grasp of the bridge construction progress is crucial to the quality assurance of the entire project. The existing construction progress management faces issues such as over‐reliance on manual inspections, difficulties in cost control, and the lack of 3D information in image identification‐based methods. Thus, this paper proposes an intelligent bridge construction progress management method combining an improved YOLOv5s model and 3D GIS. It focuses on researching key technologies, including the intelligent identification of bridge construction structures based on an improved YOLOv5s model that incorporates the RepVGG module and edge feature extraction algorithm, the 3D reconstruction of construction structures with semantic constraints, and the construction progress management based on 3D GIS. A prototype system is developed and case experiment analysis is carried out. The experiment results indicate that the proposed method in this paper can effectively identify bridge construction structures, with an average identification accuracy of 84.6%, and a 3D reconstruction accuracy of construction structures of approximately 97%. It supports 3D comparative analysis between bridge construction progress and the planned schedule, automatically generates bridge construction progress results. By making full use of multisource on‐site data to reduce management costs and enhance the intelligence level of progress management, and provides technical support for building smart construction sites for modern bridge engineering.
Fu et al. (Wed,) studied this question.