An Integrated Earthquake Damage Detection System

An integrated damage detection system, which was developed to detect the collapsed buildings due to Kocaeli earthquake, occurred on 17 August 1999, is introduced. The implementation of the system was carried out in a selected area of the city of Golcuk, TURKEY. The developed system is composed of three components: (i) input, (ii) analysis and (iii) output. In the input component, the post-event panchromatic aerial image of the study area and the vector building boundaries are fed into the analysis component. In the analysis component, an approach, which utilizes the building grey-value and the gradient orientation, is implemented. In the overall assessment of the buildings, the grey-value and the gradient orientation-based results are combined and a single label (collapsed or un-collapsed) is assigned to each building in an integrated manner. In the output component, the results of the total labeling are presented in both graphical and textual mode. The results show that of the 284 buildings analyzed, 254 were labeled correctly as collapsed and un-collapsed providing the producer’s accuracies of 81% and 92%, respectively. If the buffer zones are generated around the buildings and the assessments are carried out within the buffered building polygons, 258 buildings are labeled correctly providing the producer’s accuracies of 81% and 94%, respectively for the collapsed and un-collapsed buildings. It can be concluded that the collapsed buildings caused by the earthquake can be successfully detected from post-event aerial images using an automated system approach.