Spaceborne VHR SAR data for detection of sparse man-made features in wide-scale safety inspection after an earthquake

In the immediate aftermath of a major earthquake, inspection of structures to determine their damage level is needed to determine possible risks connected to aftershock collapses. In rural areas the presence and location of sparse man-made structures may not be readily available information, and in this context, the use of VHR SAR data may be used to detect and locate features connected with the presence of structures which require inspection. Locating urban areas and urban patches in spaceborne images is a largely explored research problem, but the focus of the proposed solution was mainly on accurately tracking urban boundaries rather than finding smallest, sparse clusters of buildings. In this paper, the authors build on their experience with the 6th April 2009, Italy earthquake to develop a detection technique focussing on this particular application. A first version of the algorithm, based on adaptive local thresholding of the reflectance data [1] has now been integrated with a more refined feature detector based on the widely experimented W-filter [2]. The enhanced method and the results of preliminary testing on pre- and post-event COSMO/SkyMed images will be presented and discussed. [1] F. Dell’Acqua, Polli, D., Ferrari, S. (2010) “Spaceborne support to post-earthquake operations: detection of sparse man-made features in wide-scale safety inspection”. Accepted for oral presentation at EUSAR 2010, Aachen, Germany, 8-10 June 2010 [2] M. Negri, P. Gamba, G. Lisini, and F. Tupin, “Junction-aware extraction and regularization of urban road networks in high-resolution SAR images,” IEEE Trans. Geosci. Remote Sens., vol.44, no.10, pp.2962-2971, Oct. 2006