Proposal of a Disrupted Road Detection Method in a Tsunami Event Using Deep Learning and Spatial Data

Tsunamis generated by undersea earthquakes can cause severe damage. It is essential to quickly assess tsunami-damaged areas take emergency measures. In this study, I employ deep learning and develop a model using aerial photographs road segment data. obtained data from the taken after Great East Japan Earthquake; used was YOLOv5. The proposed method based on YOLOv5 determine damaged roads pictures disaster. feature of use training images separated specific range distinguish presence or absence damage related tsunami. results show that more accurate than comparable traditional method, which constructed labeling areas. highest F1 score 60~78%, while 72~83%. could not detect locations where it difficult status photographs, such as houses are completely damaged. However, able them.