Recognition of damaged building using deep learning based on aerial and local photos taken after the 1995 Kobe Earthquake

Abstract

A quick and accurate method for investigating damage to buildings is required for proper disaster response. In this study, we investigated whether building damage can be grasped by applying Deep Learning, which is one of machine learning methods, to aerial and local photographs for buildings damaged by the 1995 Kobe earthquake. As a result, it became clear that buildings with severe damage can be identified with accuracy of 86.0% for aerial photographs and 83.0% for field photographs. In addition, from the photograph of the site it became clear that the collapsed building can be identified with an accuracy of 98.5%.