3D- Object Oriented Innovative Approach for Early Damage Assessment after Seismic Event
Milone, Maria Vittoria1; Mormile, Martina1; Baiocchi, Valerio1; Dominici, Donatella2
1University La Sapienza, ITALY; 2University of L'Aquila, ITALY
All the seismic events that have recently affected many areas of the planet point out the need to respond in a rapid and effective way to emergencies that may occur. With this paper we want to illustrate an attempt to combine traditional change detection (pixel-based and object-oriented) techniques with DSM extraction in order to update cartography of the areas strucked by seismic events. This innovative method can be called 3D-object oriented. Combining both techniques it's possible only when stereoscopic and multispectral images are available; stereoscopic images can be used for DSM extraction while change detection and multi-spectral classification algorithms can be applied on a monoscopic image. In this case of study we tested this innovative approach on the city of l'Aquila, in the Abruzzo region, because of its peculiar position laying partially on an ancient lake-bed that amplifies seismic activity and for this reason, a few earthquakes mark its history. For this site a high resolution Worldview 2 monoscopic image is available, on which we have tested object-oriented methodologies comparing them with the traditional pixel-based algorithms to classify roads, buildings, green areas etc, and so make a first damage assessment. This test has been completed with height information extracted from DSMs of the same area. There are several advantages in the combined use of these two techniques; in fact change detection is not always effective where new buildings are present, because, for example, ground with the same radiometry of a roof maybe classified as a building, or a collapsed building after a seismic event may have the some radiometry of an undamaged building. Possible further developments of this study provide the application of this methodologies to different and larger areas interested by seismic events.