The aging of concrete buildings raises significant safety and maintenance concerns, necessitating efficient crack detection methods. This study proposes an automated UAV-based crack detection and measurement system using deep learning and high-resolution image processing. UAVs collect images remotely, and a CNN-based YOLO model detects crack regions. These regions are enhanced using the VDSR algorithm for high-resolution transformation. Crack widths are measured via skeletonization and contour analysis, with pixel size calibrated using camera specifications. The proposed method overcomes limitations of previous approaches by enabling precise crack measurement from a safe distance. Experimental results demonstrate its ability to detect cracks as small as 0.3 mm while maintaining inspection efficiency. This integration of UAVs and AI enhances building maintenance reliability and sustainability.