Building destruction and damage assessment after earthquake using high resolution optical sensors. The case of the Gujarat earthquake of January 26, 2001

Abstract

Various methods have been developed in order to assess and detect building's damages in urban areas due to earthquakes using medium resolution optical satellite data. If large destroyed urban zone could be detected, damages at the scale of the building could not be clearly identified from the medium resolution sensors using classical change detection methods. Since the launch of the high resolution sensors such as Ikonos, QuickBird or Spot 5, experts in remote sensing have tried to apply classical methods of image analysis to those new data, in order to detect building damage. However, the low accuracy of the results have often led to the use of visual interpretation techniques. A new method based on the detection of morphological anomalies on building boundaries is experimented here, in order to detect damages at the scale of an element, using high resolution optical imagery. The image processing method is based on the comparison between the convex hulls of a building and their actual shape after the earthquake. The application was realized on the city of Bhuj (India), following Gujarat earthquake of January 2001. If only large destroyed urban zones could be detected from medium resolution, the use of this new approach on high resolution allowed us to detect damages at the scale of the building.