Oil & Gas Middle East – November 2018

(Jacob Rumans) #1

14


NEWS

NOVEMBER 2018 oilandgasmiddleeast.com

STARTUP When people see a fire,
they generally run. A brave few
might try to help. But when a
fire started near the neighbour-
hood of two young Lebanese
men, Charlie El Khoury and
Nicolas Zaatar, they decided
to create a long-term solution
using drones and artificial intel-
ligence (AI).
Originally a final year project
for university, the pair created
NAR, which is Arabic for ‘fire’,
in February 2016. The Lebanon-
based startup is creating a real-
time artificial intelligence-pow-
ered aerial inspection platform
called Raven AI.
It is not complicated: Raven
AI is a software that allows
drone cameras to understand
what they are seeing, so the
drone can detect potential risks.
While the drone is flying

Lebanese startup NAR uses drones and AI


for safety inspection


The startup’s machine learning software, Raven AI, provides real-time inspection reports


5,
The number of hours that
NAR (and Raven AI) has
spent doing inspections.

and inspecting operations, a
live feed of the footage, which
appears on a tablet through the
Raven AI application, provides
analysis of the objects it is scan-
ning. Using thermal imagery, it
can even detect leaks.
This is nothing new: the oil
and gas industry, among many
others, has been using drones
to collect information for
some time. But adding AI and
machine learning to the drone
is an innovation of the past few
years—and some hope to con-
nect drones to SCADA systems,
so they can collect asset data
wirelessly while flying over the
field.
El Khoury and Zaatar hope to
turn drones in “flying intelligent
solutions,” according to NAR’s
website. Instead of sending
drones on-site to simply collect

information, leaving the inspec-
tion team to sift through all of
the resulting photos and videos,
NAR would have drones analyse
that information in real time.
It hopes to save time and cut
costs, and the company claims
to effect an 85% reduction in
time spent generating inspec-
tion reports.
Because of the nature of the
software, which is powered by
AI and machine learning, the
more time it spends in action,
the more accurate it becomes.
So far, NAR has completed more
than 5,000 hours of inspection
and generated more than 800
inspection reports in real time.
The startup went through
Speed@BDD, a Lebanese ac-
celerator, followed by another
accelerator in Silicon Valley
called Blackbox Connect.
Free download pdf