Automated building damage assessment using remote-sensing imagery

J. Arn Womble, Kishor C. Mehta, Beverley J. Adams

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Scopus citations

Abstract

The automated comparison of before-and-after remote-sensing imagery provides an effective means for rapid and widespread assessment of windstorm damage to individual buildings. The development of automated damage-assessment algorithms involves the classification of building damage signatures from a remote-sensing perspective, the identification of corresponding temporal change metrics, and the correlation of remote-sensing change signatures with actual field-based damage observations. Hurricanes Charley (August 2004) and Ivan (September 2004) marked the first major hurricanes for which high-resolution satellite images were available. These storms provided an exceptional set of before-and- after images. Investigators from Texas Tech University and ImageCat, Inc. obtained temporal satellite image sequences and performed associated ground-truthing damage surveys for these major hurricanes. This paper chronicles the use of the before-and-after hurricane imagery to develop remote-sensing-based damage scales for various building inventories; the correlation of remote-sensing damage metrics with field-based damage investigations; and the progress in automated damage assessment using temporal image sequences.

Original languageEnglish
Title of host publicationForensic Engineering (2007)
DOIs
StatePublished - 2007
EventForensic Engineering Sessions of the 2007 Structures Congress - Long Beach, CA, United States
Duration: May 16 2007May 19 2007

Publication series

NameForensic Engineering (2007)

Conference

ConferenceForensic Engineering Sessions of the 2007 Structures Congress
CountryUnited States
CityLong Beach, CA
Period05/16/0705/19/07

Fingerprint Dive into the research topics of 'Automated building damage assessment using remote-sensing imagery'. Together they form a unique fingerprint.

Cite this