Natural disasters claim thousands of
lives annually, ripping apart communities and demolishing entire
infrastructures. In the United States alone in 2003, there were
438 weather-related fatalities, nearly 3,000 injuries and $11.4
billion in damages. Although the effects of natural disasters on
human lives aren’t easily mitigated, there are some effective
solutions to the problems of disaster planning, prevention, and
response and recovery initiatives.
One provider of such solutions is Long Beach, Calif.-based
ImageCat Inc. (www.imagecatinc.com), which develops innovative
approaches to earthquake risk management and post-disaster damage
assessment. Supported by grants from the National Science
Foundation (www.nsf.gov) and the Multidisciplinary Center for
Earthquake Engineering Research (www.mceer.buffalo.edu), ImageCat
has pushed the geospatial technology envelope when it comes to
analyzing and mitigating risks associated with natural disasters.
Since Earth Imaging Journal brought readers “Earthquake Recovery:
QuickBird Imagery of Algeria Supports Damage Detection and Relief
Efforts” (Vol. 1, No. 1—online at
www.eijournal.com/algeria.asp), ImageCat has custom-built a
robust portable notebook-based reconnaissance system that links
high-resolution QuickBird satellite imagery from Longmont,
Colo.-based DigitalGlobe (www.digitalglobe.com)
to a real-time Global Positioning System (GPS) feed. The system,
Visualizing Impacts of Earthquakes with Satellites (VIEWS), has
been used successfully in the field internationally to investigate
damage from recent deadly earthquakes and windstorms.
For several years, ImageCat researchers have investigated how
remote sensing technologies can improve response and recovery
activities following major earthquakes. The company’s milestone
research and development activities now have spiraled into
applications that support investigations of other disasters, such
as hurricanes.
As ImageCat furthers its collaborative work with other
organizations dedicated to the same cause, it will no doubt find
more ways to integrate satellite imaging technology and multimedia
techniques, providing new solutions that improve response and
recovery efforts for years to come. Consider the important
societal benefits of the VIEWS system in the following case
studies from around the globe.
In Dec. 26, 2003, a powerful earthquake with a
magnitude of 6.6 on the Richter scale struck southeastern Iran, killing
26,000 people, injuring 20,000 and leaving 60,000 homeless. The
earthquake destroyed much of the city of Bam—an estimated 90 percent of
the city’s buildings were destroyed or severely damaged, including the
old quarter and a 2,000-year-old citadel built primarily of mud brick.
For the first time, the Earthquake Engineering Research Institute (www.eeri.org)
reconnaissance team was able to use the custom-built VIEWS system to
assess urban damage caused by the Bam tremor. With high-resolution
QuickBird satellite imagery purchased by the Earthquake Engineering
Research Institute and the University of California at Irvine (www.eng.uci.edu),
providing citywide coverage, VIEWS let the reconnaissance team visualize
damage sustained by individual structures and directed them to the
hardest hit areas.
Using before-and-after QuickBird imagery, an automated texture-based
change detection algorithm was created, offering a “quick-look” damage
assessment and providing the focus for more detailed inspection of
building damage using visualization techniques. A visual comparison also
was drawn between enlarged views of the QuickBird images collected three
months prior to the earthquake and one week after.
The extremely detailed before-and-after images revealed widespread
changes to the city. In the building damage map, areas that experienced
extreme changes, such as collapsed structures, were indicated in red.
Clearly defined dwellings in the “before” image scene were indistinct
piles of debris in the “after” image.
QuickBird imagery also served as a basemap and
orientation device for response teams unfamiliar with Bam. To help users
gain and maintain their bearings, VIEWS tracked their current position
with a real-time GPS feed. As responders entered data, such as building
damage descriptions and photograph ID numbers, the information was
linked to their current GPS locations. The system also provided easy
recall for observations made in the field. Back in the office, VIEWS
datasets were readily transferred to a geographic information system
(GIS) environment for further analysis.
“The ability to display all the sources of graphical field data—digital
satellite images along with video or still photographs—when we were back
in the office re-created the environment for us,” explains Charles Huyck,
ImageCat’s senior vice president. “Combining these technologies was
invaluable for understanding what was happening in the field.”
During the months following the Bam reconnaissance effort, ImageCat
further honed its data collection techniques and prepared for another
potential deployment.
Hurricane Charley struck the Florida coast on
Aug. 13, 2004. The most severe windstorm to strike the United States
since 1992, Hurricane Charley incited President Bush to issue a
presidential disaster declaration for 25 counties in the southeastern
region of the United States. The Category 4 hurricane resulted in the
loss of at least 27 lives, and caused more than $15.4 billion of damage.
After a disaster such as a hurricane, clean-up operations typically are
initiated as soon as possible. There is a narrow time window for
documenting perishable damage to buildings and infrastructures. The
ImageCat team was quickly on the scene with newly collected satellite
imagery and the VIEWS system.
“The VIEWS program was built specifically around timely satellite imagery,” says Beverley Adams, remote
sensing group leader at ImageCat. “It’s designed for
reconnaissance teams to take into the field and start recording
on-the-scene information immediately, so the fresh imagery really
supported our efforts after Hurricane Charley.” VIEWS was used to
collect perishable damage data on two separate field trips that took
place between Aug. 18-21, 2004, and Aug. 24-27, 2004. Efforts focused on
the towns of Port Charlotte and Punta Gorda. The response team used
QuickBird imagery collected before the hurricane—on March 23, 2004—and
in the immediate aftermath—on Aug. 14 and 19, 2004.
For the post-Hurricane Charley assessment,
ImageCat acquired funding from the National Science Foundation’s Small
Grants for Exploratory Research program, the Natural Hazards Research
and Applications Information Center Quick Response program and the
Multidisciplinary Center for Earthquake Engineering Research
post-disaster field deployment program. The Wind Science and Engineering
Research Center at Texas Tech University also accompanied ImageCat on
its field reconnaissance trips.
The field trips enabled the analysis team to validate building damage
characteristics identified on the satellite imagery. Traditional methods
of post-hurricane damage assessment involve walking surveys where damage
indicators, such as the failure of roof covers, roof structures and
windows and doors, are logged on a spreadsheet.
“Engineers typically have to manually record what they see on a
clipboard,” relates Adams. “This is very time consuming and slows down
the data recording process.” By contrast, the ImageCat team was equipped
with satellite imaging technology and the VIEWS field reconnaissance
system, which substantially accelerated and streamlined data collection.
Combined with a digital camera and digital video recorder technology,
VIEWS can be used from either a moving vehicle or on foot during a
walking tour. “Our technology-driven approach has increased our
efficiency 25-fold,” says Adams. “Traditional survey techniques
typically enabled 20 to 100 buildings to be surveyed in one day. Using
VIEWS and satellite imagery, our field experts were able to capture
damage data for an average of 2,500 buildings per day.” Capturing the
data digitally also produced a permanent visual record of damage sustained by individual
structures.
Working with the Multidisciplinary Center for Earthquake Engineering
Research, ImageCat took its reconnaissance efforts one step further with
the development of tandem Internet- and desktop-based virtual
reconnaissance systems (VRSs), which enable ImageCat analysts to
integrate, share, visualize and analyze post-disaster field data
collected using VIEWS. After Hurricane Charley, VRS provided researchers
with easy access to the satellite imagery, GPS readings and
geo-referenced video and photographic records for Port Charlotte and
Punta Gorda. Users can toggle between multitemporal and multisource
satellite images, and explore these images in detail using zoom and pan
functions. By overlaying the data with GPS routes and selecting a GPS
point, users can view corresponding video footage and scroll through the
archive of nearby photographs.
In earthquake hit Niigata, on the Sea of Japan’s
coast, on Oct. 23, 2004. With a magnitude of 6.8, it was the deadliest
to hit Japan since the Kobe earthquake killed more than 6,400 people in
1995. The Niigata tremor left 36 dead and more than 85,000 people
evacuated or homeless. More than 5,700 homes and 2,887 buildings were
damaged; 312 homes were destroyed. Hundreds of aftershocks throughout
neighboring regions buckled highways, twisted rail lines, collapsed
bridges and triggered landslides.
ImageCat’s Charles Huyck traveled to Niigata with the Earthquake
Engineering Research Institute reconnaissance team to record the
aftermath of the earthquake. “Because VIEWS can be deployed by car as
well as on foot, large swaths of land were rapidly videotaped during
that narrow window of time between the devastating event and the onset
of clean-up operations,” relates Huyck.
In the georeferenced video taken in the Niigata region, green, yellow or
red building tags, indicating building safety, could be cross-referenced
with GIS data such as soil or ground motion.
“Landslides are one of the biggest worries in the Niigata region, which
has been experiencing heavy rainfall since the earthquake,” adds Huyck.
“The QuickBird imagery collected afterward provided an excellent view of
the extensive landslides in areas that were often inaccessible by land.”