Advanced UAV photogrammetry for precision 3D modeling in GPS denied inaccessible tunnels

Unmanned Aerial Vehicles (UAVs) have evolved into potent tools for both researchers and professionals. Their use has expanded significantly in recent years across diverse fields of research and engineering, primarily owing to their image-capturing capabilities. These capabilities offer benefits such as time efficiency, cost-effectiveness, minimal fieldwork, and high precision. This study focuses on harnessing UAVs’ photogrammetric technology to create a reliable and accurate 3D modeling process. By doing so, we assess the feasibility of modeling and mapping hard-to-access tunnels using imagery acquired from UAVs combined with photogrammetry. This approach consistently provided a faithful representation of the environment under study. Three distinct experiments underpin this research. The first aims to construct a 3D model using a singular directional light source, evaluating its precision in a concrete tunnel setting. The second experiment involves crafting a 3D model of an underground sewer tunnel using both unidirectional and omnidirectional light sources. The final experiment leverages the insights from the preceding tests to create a 3D model via a specially-designed UAV system, gauging the accuracy of this 3D photogrammetric model within what is termed in this study as a "pedestrian tunnel." However, there are inherent challenges to this approach. For instance, augmenting a drone with additional equipment changes its weight distribution, potentially compromising its stability. An improperly balanced payload could result in uneven flight patterns, rendering the UAV harder to manage.