Monitoring Post-Earthquake Search and Rescue Operations through UAVs Vision: Teams, Equipment and Structural Damage Detection

With the development of unmanned aerial vehicles and deep learning-based computer vision algorithms, there is an increasing interest in using these technologies together in managing pre- and post-natural disaster operations which are serious crises. This study focuses on enhancing search and rescue efforts and crisis management, particularly in time-critical and high-accuracy tasks, especially for earthquakes. It addresses the detection of teams, equipment, and structural damage by analyzing a dataset comprising images captured by UAVs from various lighting conditions, altitudes, and viewing angles using YOLO algorithms. Comparative results between the latest YOLOv8 algorithm and YOLOv5 are presented, showing an accuracy rate of 88% mAP, along with high speed, which will enable critical tasks in search and rescue operations to be carried out effectively with sufficient equipment and accurate identification of needs.