Case Study: Rapid IA & PA damage assessments in a cloud environment
Track: Disaster and Emergency Management
Authors: Scott Perkins
Title: A Case Study using commercial airborne imagery to perform rapid IA & PA damage assessments in less than 24 hours in a cloud computing environment.
In 2013 the Department of Homeland Security made awards to four contractors to provide Remote Sensing to Support Incident Management and Homeland Security. The services under this contract may be performed in any of the fifty United States or its territories, for government agencies under contract ##HSHQDC-13-D-RS001.
The primary requirement of these contracts is to acquire fresh airborne imagery or LiDAR and upload onto FEMA’s geospatial portal fully GIS useable E/O imagery or LiDAR within 48 hours from receipt of Notice to Proceed (NTP). In June of 2013 all four contractors received an initial tasking, an ‘emergency response exercise’ to test their capabilities to meet or exceed the 48 hour delivery requirement.
After the ‘exercise’ in follow on meetings with FEMA and DHS, Geospatial Management Office (GMO) within the Office of the Chief Information Officer (OCIO) a need was identified to provide very rapid IA & PA damage assessments using commercial airborne imagery. The goal was identified as providing the IA & PA damage assessment as soon as possible, even in advance of the ortho corrected imagery if possible.
Granbury, Texas. May 16, 2013 EF-4 Tornado
Aero-Metric collected 4-band E/O imagery using a large format DMC sensor of the impacted area. Imagery was ortho corrected and image tiles created referenced to the US National Grid within 24 hours after acquisition, available on Areo-Metric GeoAPP cloud based service.
Using the NIR band an IA (individual assistance) damage assessment was performed using FEMA’s protocol identifying four (4) levels of damage. The rapid IA was completed in less than 4 hours.
The results of the Granbury, TX case study on rapid damage assessment clearly demonstrated that rapid IA or PA damage assessments can be performed in less than 24 hours from acquisition of airborne commercial imagery and can be fully GIS useable via a GeoAPP or Geospatial portal in advance of the delivery of the ortho corrected image tiles.
Further study is planned to test tools, techniques and processes for automation of the IA & PA damage assessment task. Of particular interest is the potential of utilizing algorithms developed for pattern recognition applications in the biometrics field for identification of debris, roof damage and other damage assessment applications.