Accuracy and Efficiency of UAV-Derived 3D Models for Building Inspection: Insights from Kompleks Sukan UiTM Perak Branch

Abstract: This study explores the use of unmanned aerial vehicles (UAVs) for detecting building defects through the generation of accurate three-dimensional (3D) models. It specifically focuses on identifying flaws in hard-to-access areas and comparing measurements between real-world structures and their 3D representations. Traditional inspection methods for complex structures are often hazardous, time-consuming, and limited by accessibility. In contrast, UAV-based photogrammetry offers a safer, more efficient alternative for comprehensive visual assessments. High-resolution images of the Kompleks Sukan building at UiTM Perak Branch were captured using UAVs and processed with Pix4D Mapper to produce detailed 3D models. These were then compared against conventional measurement techniques. Results showed a high degree of accuracy and consistency, with only minor discrepancies. The models also revealed defects that would be difficult to detect using traditional methods, highlighting the advantage of UAVs in reaching inaccessible zones. Root Mean Square Error (RMSE) values further confirmed the accuracy of UAVs-based photogrammetry. Overall, the study demonstrates the effectiveness of UAVs in enhancing building inspection processes by improving safety, saving time, and providing reliable structural data. The findings support the broader adoption of UAVs technology in the construction and maintenance sectors for informed, data-driven decision-making.