After two eartquakes occurred in Kahramanmaraş on 6 February 2023, more than 50,000 people were lost their lives and buildings were damaged in various cities. Building damage assessment necessity arose and many buildings were photopraphed accordingly. Within the scope of this study, it is aimed to do damage classification (collapsed or not collapsed) of buildings in colored images taken from the side using deep learning techniques. The model using transfer learning achieved 88.5% test accuracy.
Keywords — image classification, post-eartquake building damage, deep learning, transfer learning.