Korean Journal of Computational Design and Engineering

UAVs and AI technologies have become innovative tools for efficiently diagnosing building facades. UAVs collect millimeter-scale crack data, and AI analyzes this information to detect structural defects. However, practical limitations exist. Diagnosing such fine cracks requires UAVs to fly very close to building surfaces, which is challenging due to environmental factors like irregular airflows and weak satellite signals, making precise control difficult. These issues hinder the practical application of UAV-based diagnostics. To address these challenges, this study proposes a crack diagnosis method using high-resolution algorithms. These algorithms enhance image data captured from a distance to match the quality of close-range data. This allows UAVs to scan large facade areas from a safer distance while maintaining the precision of close-up inspections. Consequently, this approach overcomes existing limitations, enabling more efficient and safer diagnostic processes.

키워드 Unmanned aerial vehicles, High resolution, CNN, Crack, Building Inspection