Complete Perimeter Solution
Radar and deep learning technologies finally come together
- By Yaron Zussman
- Oct 01, 2018
An alarm is triggered in the
middle of the night at a remote
power plant. An offsite
security patrol officer receives
the alert via the video
management system and springs into action,
driving more than 30 minutes to the remote
location in hopes of mitigating the threat before
the situation escalates. Upon arrival, the
security guard realizes the alert was simply
an animal at the fence line that triggered a
nuisance alarm. This scenario plays out all
too often, whether it be at a utility substation
or commercial business. False alarms
are detrimental to security efforts, wasting
both time and money.
This is just one of many instances where
artificial intelligence (AI) and deep learning
have become critical for security solutions to
function optimally. While practically everyone
in the security industry has been talking
about AI and its possibilities, some are still
unclear on how this technology works and
what it has to offer.
A key trend for 2018 has been the integration
of AI with various security systems.
AI-enabled surveillance cameras, facial recognition
software, and fingerprint authentication
scanners are just a few examples.
One unique AI technology pairing gaining
traction in the perimeter security sector is
radar. Thanks to its superior reliability and
wide field of coverage, radar has emerged as
a prime candidate for integration. Taking the
most dependable radar sensors and optical
sensors and merging these with deep learning
programs via advanced video management
software, enables higher-performing
analytics and enhanced perimeter security.
Before discussing this dynamic technology
combination, let’s look at radar and its uses
for the security industry.
Radar Technology
Radar stands for “Radio Detection and
Ranging,” and uses radio frequency waves to
identify and track targets. In simple terms,
radar is an object detection system that sends
out radio waves to determine the range, direction,
speed, and altitude of an object.
Unlike video or thermal imaging cameras,
radar performance is not adversely affected
by weather, lighting, or other external forces.
Radar provides continuous coverage of any
protected site in all weather and lighting conditions,
allowing for volumetric perimeter
protection while keeping power consumption
to a minimum.
Radar has been commonly used in government
and military applications, giving the
misconception that it is only meant for highlevel
deployments. However, radar’s wide
coverage capabilities make it an ideal system
for a variety of perimeter applications including,
data centers, oil and gas refineries,
logistic centers and education campuses.
With the ability to reach up to 120-degrees
in azimuth and 30-degrees in elevation,
a single radar is able to optimally cover an
area despite changing topography. When radar
is placed around the perimeter of a site,
targets that approach the “fence line” will
be automatically detected and tracked. For
security forces using traditional video cameras
alone, detecting humans or vehicles at
far ranges can be near impossible, allowing
threats to come dangerously close to protected
assets. On the other hand, radar can
detect humans or vehicles up to 1,000 meters
away, alerting security forces long before
they reach the boundary.
Integrating this technology with PTZ
cameras enables slew-to-cue capabilities, allowing
security personnel to have eyes on
exactly what the radar is tracking for maximum
situational awareness. New technology
advancements have made radar even more
effective. Here’s an overview behind the AI
technology driving this enhancement.
AI and Deep Learning
The lines between artificial intelligence and
deep learning can be unclear for many consumers.
The easiest way to think of these
technologies is in concentric circles, with AI
being the largest or the umbrella concept,
and deep learning in the center as it is the
technology that enables AI’s abilities. Artificial
intelligence refers to systems that perform
tasks that identify patterns objects. In
security, AI capabilities translate to greater
data analysis and business intelligence.
For example, AI security systems can now
detect and even recognize faces stored in a database,
adding a whole new layer to duel-authentication
access control and management. License plate recognition is another common AI function.
Deep learning is the recent breakthrough within the field of
AI that allows for machines to classify objects and make decisions
based upon what they have “learned.” Deep learning algorithms that
are already being deployed can not only enable AI devices to learn
and perceive their environments but can also learn to differentiate
everyday occurrences from abnormalities. This pattern recognition
may seem simple, but it is a huge victory for analytics software. By
giving machines “brains” to match their “eyes,” these offerings allow
for higher levels of accuracy and reliability in both object and
behavior classification.
The Merging of Radar and Deep Learning
While still a relatively new integration, the concept of enhancing perimeter
systems, featuring radar and other video technologies, with
deep learning is taking hold in the security industry. The idea is simple,
combine the best technology for target detection with the best
technology for target classification and merge them to create a fused
engine that yields the lowest possible number of false alarms. Some
AI software work only off of a radar signal, but more advanced solutions
are extending the AI to the accompanying video stream to
analyze data from both sources to get the most accurate results.
In the latter scenario, it all begins with radar detection. Once the
radar filters out false alerts, it sends the validated target tracks detects
to the VMS, which in turn, cues an integrated PTZ camera to follow
the movements of said target. As the PTZ camera follows the intruder,
deep learning software analyzes the video stream and tracked movement
from the radar to classify the object. Once the target has been
classified and validated, which takes less than a single second, a verified
alert is generated, and the system will log and record the event.
To better illustrate the point, take a large tech company with data
centers in various geographic areas. One of the greatest challenges
faced by these facilities is the fact that many are in remote areas with
wildlife surrounding the perimeter. While the complexes must be protected
from vandalism and physical intrusion, these sites are regularly
plagued by nuisance alarms. The largest companies own tens of data
centers. Even if the number of daily false alarms is below five, the
company could receive dozens a day from all the data centers combined.
This inundates security systems.
Many times, security guards or law enforcement will end up responding
to erroneous alarms, which take their attention away from
true threats. By utilizing a deep learning target classifier that analyzes
data from both the radar and video stream, the nuisance alarm rate
(NAR) can be significantly reduced. In fact, for successful deployments,
the system can replace physical guards, heavily reducing security
operating costs.
There is a clear demand for perimeter security solutions that are
effective in all conditions, and that reduce the NAR. While the integration
of radar, video and deep learning technologies are still in
its infancy, development has been rapid and there is no doubt that
this will continue. This new integrated solution
enables security to run at maximum efficiency
while keeping costs minimal, simply by marrying
the most reliable sensors with the most advanced
analytics software available.
This article originally appeared in the October 2018 issue of Security Today.