BIM modeling from UAV imagery for transportation infrastructure and visualization on a GIS platform

Abstract

Integrating digital technologies such as UAV, BIM, and GIS opens new approaches for surveying, managing, and maintaining transportation infrastructure in the context of digital transformation in the infrastructure sector. This study proposes a three-step integrated workflow comprising (i) the acquisition and processing of UAV imagery to generate a point cloud, (ii) BIM modeling based on the point cloud, and (iii) the conversion and integration of the BIM model into a GIS environment for infrastructure visualization and information querying. The experiment was conducted on a segment of 1K Highway in Binh Duong Province, utilizing a DJI Mavic 3 Enterprise drone and Agisoft Metashape software to process images using Structure-from-Motion (SfM) and Multi-View Stereo (MVS) techniques. The resulting point cloud consists of over 173 million points, with georeferencing errors (GCP adjustment) ranging from 2 to 11 cm. The BIM model was developed in the Revit environment, coordinate-synchronized by using Civil 3D, and then exported to IFC format for integration into the ArcGIS Pro platform. The results demonstrate that the model accurately reflects the road geometry and auxiliary components, effectively supporting spatial analysis and infrastructure monitoring tasks. This study also proposes a future development direction: linking the model with IoT sensors to establish a real-time intelligent infrastructure management system.