Improvement of Virtual Tour Depiction Potential Utilizing Active and Passive Sensing Data Fusion

Abstract

The generation of geospatial information about objects has become a well-established field of study, facilitated by the advent of diverse techniques and technologies. Furthermore, the demand for more realistic and detailed three-dimensional (3D) models and precise digital twins has led to a surge in interest in 3D modeling. The integration of high-quality 3D models and high-resolution textures in virtual reality (VR) has constituted a well-established field of study within this context. The unmanned aerial vehicle (UAV) is a highly sought-after technology due to its ability to facilitate high-resolution aerial photo acquisition in a relatively short period of time, at a relatively low cost, and with minimal labor. In addition, terrestrial laser scanning (TLS) is a widely utilized technology due to its competence in delivering high accuracy and precise 3D point clouds. Nevertheless, both technologies are subject to inherent limitations in 3D reconstruction. The UAV technique is based on the passive remote sensing principle, which presents certain challenges when attempting to view objects in areas that are partially or fully occluded, including the forest understory and roofed structures. Furthermore, the TLS technique, which is based on active remote sensing, encounters challenges in capturing the upper parts of objects, limited mobility, and high operational costs. In this study, a virtual tour of the Gebze Technical University (GTU) campus was created by using UAV and TLS data. Subsequently, the virtual tour depiction potential was improved through active and passive sensing data fusion and 3D models generated from UAV data were evaluated in comparison with 3D models produced through data fusion. In conclusion, the limitations of the applied techniques in enhancing the potential of the virtual tour to represent objects were examined in terms of their advantages and the impact of object geometry on their effectiveness. © 2025 Elsevier B.V., All rights reserved.