Automated assessment of building vulnerability with high resolution IKONOS data for Padang, Indonesia

Abstract

The primary objective of this study is the development of an automated process to derive building vulnerability towards tsunami impacts based on an object-based extraction approach using very high resolution (VHR) satellite data. The developed automated process allows transferability to other endangered urban areas for the derivation of spatially distributed information on physical vulnerability indicators (e.g. built environment vulnerability). The city of Padang, Indonesia, with its highly structured endangered urban area, was chosen as test site. High resolution data from the IKONOS sensor were used for this study. First, an object-oriented urban land cover classification was developed. The accuracy assessment resulted in more than 80% correct detected houses. The capabilities of remote sensing enable up-to-date and area-wide spatial information, a promising cost-effective approach for the challenges of managing urban areas. Based on tsunami post disaster field surveys related to building damages conducted by Tsunarisque Project team in the Aceh region after the 26th December 2004 event as well as a field surveys in Padang, building vulnerability classes were defined. The built-up mask of the urban land cover classification was used to assign the defined building vulnerability classes on the whole scene, using spectral information as well as the geometry of features to differentiate those building vulnerability types. The presented study delivers important information on physical vulnerability indicators of the built environment and infrastructure on a high spatial resolution. In combination with modelled tsunami propagation scenarios on land, detailed risk assessment information is obtained and may deliver useful insights for designing disaster risk reduction strategies.