718
August 2014
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
It is mid-morning in New Guinea during this early July
day. The year is 1937. Amelia Earhart and her navigator
Fred Noonan are departing from the city of Lae for the next
leg of their journey around the world
.
As they are in transit
over the central Pacific Ocean they make their last radio
communication. It is reported that the plane never arrived at
their destination on Howland Island. For the next few weeks
the U.S. Navy conducted a sea and air search over 250,000
square miles. As the disappearance has remained a mystery
for years, no remnants of the plane, Ms. Earhart, or Mr.
Noonan were ever found in the mid-Pacific Ocean region. Fast
forward almost 77 years later to March 2014 and Malaysia
Airlines Flight 370 disappears during flight. With social media
and satellite technology, many people from around the world
(even the author of this column) felt a connection to help and
use Tomnod from DigitalGlobe to assist in the search to bring
closure to the families of the passengers. Several countries
tasked their satellites and naval resources in the search. To
this date no location has been identified in the search of the
plane or its passengers.
There was hope in the two months after the tragic
disappearance of Flight 370 that the latest in social media
concepts and remote sensing technology could lead us to an
answer. “We have reached a point in time where we can task
satellites and analyze imagery within minutes to hours after
an event occurs on the planet”, stated Luke Barrington (Senior
Manager, DigitalGlobe’s crowdsourcing platform, Tomnod).
Barrington stated that even though Malaysia Airlines Flight
370 was not found with the Tomnod search, the eight million
people who tagged more than 15 million satellite image clues
from over one million square kilometers helped to rule out
possible locations where the plane could have crashed.
Tomnod was founded out of a yearning for exploration and
a curiosity to search for the 800 year old tomb of the Mongol
Emperor, Genghis Khan. In a National Geographic Society
(NGS) Exploration led by Dr. Albert Lin, a Research Scientist
at the University of California at San Diego (UCSD), the Valley
of the Khans Project helped to initiate the early development
of crowdsourcing of satellite imagery. News about this NGS
Exploration reached the DigitalGlobe Foundation who offered
to provide satellite imagery to assist in the search. Around this
time Albert Lin had begun collaborating with fellow graduate
students at UCSD, Luke Barrington, Shay Har-Noy and Nate
Ricklin. The four colleagues quickly realized the magnitude of
The Advent of Geospat ial Crowdsourcing
visually reviewing the vast amounts of imagery coming from
DigitalGlobe. Barrington coincidently was working on his PhD
research using crowdsourcing to analyze massive photo and
music data sets. This crowdsourcing research inspired what is
now Tomnod. The team shared the DigitalGlobe imagery on a
National Geographic web interface and the first crowdsourcing
campaign began, tasking people from around the globe to help
guide the explorers in their search for the tomb of Genghis
Khan.
In deciding on a new name for their start-up company
specializing in geospatial crowdsourcing, Lin, Barrington,
Har-Noy and Ricklin chose “Tomnod”, a Mongolian phrase
meaning “Big Eye”. Tomnod’s vision is to use the “big eyes” of
everyone on the planet to analyze imagery. In 2013 Tomnod
was acquired by DigitalGlobe. Now the Tomnod platform
deploys at least one to two campaigns a week (540 to date)
on their website
) and uses their Facebook
page (facebook.com/tomnod) and Twitter account (@tomnod)
to engage people to volunteer in a variety of campaigns that
include identifying damage after natural disasters, detecting
important and changing infrastructure across the globe and
searching for missing planes, ships and people.
“You may think that people who volunteer only have an
interest in looking at imagery but the audience that joins
Tomnod Campaigns are genuinely interested in the topic at
hand and have a desire to contribute to solving the issue,”
stated Barrington. Barrington goes on to explain that for
decades we have had computers crunching numbers to solve
geospatial problems and now we are realizing we need human
interaction to make visual conclusions for some of these
tougher to solve queries. Through the use of interpretative
insight, only humans can draw on past knowledge to finalize
the analysis. This visual insight is the technique needed to
make geospatial crowdsourcing a success.
By definition, crowdsourcing is the use of many people to
solve an issue. Commercial soft drink companies use passive
crowdsourcing as they perform random data mining on Twitter
of hashtag names to analyze product trends. Tomnod uses the
concept of active crowdsourcing. The active method provides
for a controlled system as you are trying to solve a known
problem and you are targeting a specific crowd. That specific
crowd could be a group of people near natural or man-made
disasters or people like me, geospatial imagery geeks who like
looking at satellite imagery and solving real world issues.
