After a seismic event a rapid and accurate evaluation of the impact of the damages is extremely important. Such evaluation may support rescue team operations and identify the actual dimensions of the event and its potential impact on the territory and on the population. The use of Earth Observation (EO) data has been significantly increasing in the last years, particularly the use of Very High Resolution (VHR) optical images, which are able to provide detailed information at single building level. However, most of the existing approaches mainly rely on the use of remote sensing data, either optical or SAR (Synthetic Aperture Radar), and perform a classification based on change detection techniques. In this work we aim at creating a flexible tool that is able to perform a damage classification taking into account, not only EO available data, but also additional information that is supposed to be available even before the occurrence of any seismic event (a-priori data). This data includes soil vulnerability, which can play a very important role on local amplification effects as well as structural information of the individual building. Such approach, pursued within the framework of the EC-FP7 funded project APhoRISM (Advanced Procedures for Volcanic and Seismic Monitoring- grant agreement n. 606738) aims at generating maps of damage caused by a seism using both satellite remote sensing data (SAR and/or optical sensors) and ground and structural data. The basic idea is to integrate both satellite remote sensing data (SAR and/or optical sensors) with structural and ground data to improve the accuracy and limit false alarms that derive by the use of EO data only. In order to do this, we first review the general approach and methods to data fusion and we identify what is the level of information that is better to merge referring to our goals. We also examine how the structural information is evaluated and we then focus on the description of Bayesian approaches and, more specifically, of Bayesian networks. Such type of graphical approach for our data fusion tool is implemented to assess post-earthquake building damage. We validate our Bayesian networks against the real test case based on L’ Aquila (Italy) earthquake which took place on April 6, 2009. In this case, we have a set of data available to build the Ground Truth validation test set. For what concerns remote sensing data, for this event, both COSMO-Skymed Radar and Quickbird VHR optical sensors were available thus allowing a complete remote sensing dataset. The in-situ information, though fragmentary, was built using data coming from different sources, mainly from INGV (Italian Geophysical and Volcano Institute) and the Italian Civil Protection Department. The promising results of different Bayesian networks are presented showing the step-by-step approach adopted, which aims at generalising the methodology in order to further implement the network in future cases.

Earthquake damage analysis and mapping with the use of satellite remote sensing / Scalia, Tanya. - (2017 Feb 17).

Earthquake damage analysis and mapping with the use of satellite remote sensing

SCALIA, TANYA
17/02/2017

Abstract

After a seismic event a rapid and accurate evaluation of the impact of the damages is extremely important. Such evaluation may support rescue team operations and identify the actual dimensions of the event and its potential impact on the territory and on the population. The use of Earth Observation (EO) data has been significantly increasing in the last years, particularly the use of Very High Resolution (VHR) optical images, which are able to provide detailed information at single building level. However, most of the existing approaches mainly rely on the use of remote sensing data, either optical or SAR (Synthetic Aperture Radar), and perform a classification based on change detection techniques. In this work we aim at creating a flexible tool that is able to perform a damage classification taking into account, not only EO available data, but also additional information that is supposed to be available even before the occurrence of any seismic event (a-priori data). This data includes soil vulnerability, which can play a very important role on local amplification effects as well as structural information of the individual building. Such approach, pursued within the framework of the EC-FP7 funded project APhoRISM (Advanced Procedures for Volcanic and Seismic Monitoring- grant agreement n. 606738) aims at generating maps of damage caused by a seism using both satellite remote sensing data (SAR and/or optical sensors) and ground and structural data. The basic idea is to integrate both satellite remote sensing data (SAR and/or optical sensors) with structural and ground data to improve the accuracy and limit false alarms that derive by the use of EO data only. In order to do this, we first review the general approach and methods to data fusion and we identify what is the level of information that is better to merge referring to our goals. We also examine how the structural information is evaluated and we then focus on the description of Bayesian approaches and, more specifically, of Bayesian networks. Such type of graphical approach for our data fusion tool is implemented to assess post-earthquake building damage. We validate our Bayesian networks against the real test case based on L’ Aquila (Italy) earthquake which took place on April 6, 2009. In this case, we have a set of data available to build the Ground Truth validation test set. For what concerns remote sensing data, for this event, both COSMO-Skymed Radar and Quickbird VHR optical sensors were available thus allowing a complete remote sensing dataset. The in-situ information, though fragmentary, was built using data coming from different sources, mainly from INGV (Italian Geophysical and Volcano Institute) and the Italian Civil Protection Department. The promising results of different Bayesian networks are presented showing the step-by-step approach adopted, which aims at generalising the methodology in order to further implement the network in future cases.
17-feb-2017
File allegati a questo prodotto
File Dimensione Formato  
Tesi dottorato Scalia

accesso aperto

Tipologia: Tesi di dottorato
Licenza: Creative commons
Dimensione 3.02 MB
Formato Adobe PDF
3.02 MB Adobe PDF

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11573/937634
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact