Geospatial Analysis for Identifying Blocked Roads after Earthquakes: A Case Study from the Kahramanmaras Earthquake

Rapid identification of road blockages is essential for effective emergency response in the aftermath of large-scale urban disasters. This study introduces a geospatial methodology for detecting roads obstructed by earthquake-induced building collapses, utilizing UAV imagery collected after the February 6th Kahramanmaras earthquakes. A deep learning-based object detection model (YOLOv9-C) was employed to identify collapsed structures, and road blockages were inferred through intersection analysis. A total of 21.55 km of blocked roads were identified across Nurdagi, Kirikhan, Iskenderun, and Onikisubat. To assess the model’s applicability not only within seen data but also in previously unseen regions, two distinct evaluation approaches were adopted. First, Randomly Split Analysis (RSA), involved randomly partitioning the dataset into training, validation, and testing subsets. Second, Nurdagi Test Analysis (NTA), excluded all data from Nurdagi during training and used it solely for testing, in order to evaluate the model’s generalization capability in entirely untrained areas. Blocked road detection analysis, conducted using results from the RSA and NTA, yielded precision values of 0.963 and 0.961, respectively. These results demonstrate the robustness and generalizability of the proposed approach, highlighting its potential for real-time deployment in post-disaster environments where road accessibility information is unavailable or incomplete.