Efficient Procedure of Building University Campus Models for Digital Twin Simulation

Realistic digital geographical models of real-world locations are a necessary starting point for digital twin applications, especially for simulation and visualization. However, the visual fidelity of this first step is often neglected, since the effort involved is counterproductive to the main research focus. In this paper, we explore different tools and resources to assemble an efficient and convenient procedure to create a 3D digital model of a university campus that can support digital twin applications. Specifically, the terrain, buildings, and road network are combined into the CARLA project on Unreal Engine, enabling computer vision, traffic simulation, and autonomous driving experimentation.