Solving Problems
Edge-based analytics may be the solution to the problem
- By Alex Walthers
- Oct 01, 2018
The industry is changing so
rapidly that the following
statement might seem bold
but it’s true. No one wants to
buy a camera. The customer
may need a camera, they may be tasked with
retrofitting their entire security system with
cameras, but what they’re really looking for
is a solution to a problem. The camera is
only one piece of the solution, which also includes
a VMS, storage, switches, cabling, etc.
There is no one camera that will fit every
application. The customer may want to
monitor a perimeter, a parking lot, cash registers,
or any number of other areas in their
environment—each with its own unique set
of variables that determine what the solution
needs to be.
As we see camera specifications like megapixels
and frames per second (FPS) continue
to increase, it should be noted that there are
limitations to how much performance can be
squeezed onto a network. Due to bandwidth
and storage limitations of an end user’s surveillance
system, it would be difficult for
cameras to transmit 100-megapixel video at
200 frames per second. To improve system
performance, there are areas beyond these
technologies to determine new functionality
and add value to the edge.
One way to do that is through the use
of video analytics. Analytics not only adds
intelligence to cameras to make even better
use of technologies like resolution and FPS,
but they bring a more practical and effective
value to the edge and this is when solutions
are created.
Edge Applications
There are two ways video analytics can be
deployed, either within the camera at the
edge or at the server where video is stored.
If we’re talking purely about bandwidth and
storage savings, edge-based analytics make
the most sense. Similar to a smartphone,
these applications allow users to customize
their systems and add specific functionality
as needed. There’s also no wasted data. The
camera in this scenario acts as a computer,
processing the data and only sending the
data you need across the network.
For example, rather than adding more
cameras, servers or other hardware to generate
intelligence, users could add a perimeter
analytic to a thermal camera or add a license
plate recognition application to a parking
lot camera. In both examples, it would allow
the customer to add more value to their solution
and tailor the surveillance system to
their needs. This now adds intelligence to the
existing camera at their site and solves the
customer’s problem all at once.
Bandwidth and Storage
Deploying analytics applications at the edge
eliminates the need to send all the video over
the network, which saves on both bandwidth
and storage. Instead, all processing happens
at the camera allowing just data and video
clips to be sent to the head end. This can reduce
the complexity and cost of the system
by avoiding the need to install additional
servers and other connected hardware for
processing and data transfer.
The alternative server-based approach
would mean sending all video, 24/7, 365 days
a year, to a central server where it must be
processed and stored, eating up significant
bandwidth and storage space. This type of
deployment would likely require additional
servers for each individual application. This
is not only cost-prohibitive, especially for
scenes where there isn’t a lot of activity, but
this makes any event requiring investigation
more complex and time consuming.
With edge-based applications, the only
video that needs to be sent to the head end
would be a short video clip following the
activity that triggers the application. Some
applications might only require data be sent
across the network as well—without the accompanying
video. This substantially cuts
down on the amount of bandwidth you’re
sending across the network and less storage
being taken up centrally.
Beyond Security
Another benefit of edge-based analytics applications
is the ability to easily utilize video
for broader purposes that include, but extend
beyond, traditional security. A good example
of this would be a retail environment, where
there are likely to be cameras installed at
entrances throughout the store and above
cash registers. Monitoring people coming
in, whether they are stealing anything, and
keeping an eye on POS transactions may all
be done for security purposes. Edge-based
analytics can be installed to create crossfunctional
video, expanding the use case for
the video into other areas.
The camera at the front of the store could
be configured with an application to collect
and monitor demographic data (age, gender,
etc.) to collect information on a store’s target
audience and other operational data. Cameras
can also employ heat mapping analytics
to determine what areas of the store people
tend to go to, how long they linger and other
factors that could be used for marketing and
merchandising. By combining demographic
and store traffic information, a retailer could
identify a prime location for a brand of
product and then charge the manufacturer
for that premium advertising space. The retailer
could also use information gathered
from edge-based analytics to change the configuration
of the store to alleviate congestion
in a particular area.
Existing cameras could also be used to
improve customer service. Retailers are focused
on the customer experience and no
one likes to wait in a long, slow-moving line.
Cameras near checkout could run a queueline application that would alert management when lines are long,
allowing them to open another register to clear the congestion.
These are just some of the many ways edge-based analytics applications
can contribute to greater business intelligence. While a one
percent reduction in shrink is a pretty good number, a one percent
increase in sales is typically a much larger number. With this in mind,
end users and integrators may have access to operational, marketing,
advertising and other budgets beyond just security. Working with
multiple departments to meet particular needs positions installers to
offer more services, and more customization with edge-based video
analytics, which helps create long-term customers and consistent revenue
streams.
From Deterrence to Proactive Security
Initially, surveillance cameras were used mainly for deterrence, and
they were bulky, so they could be seen easily. The thought being that
highly visible cameras would deter or prevent someone from committing
a crime or doing anything malicious. Under this model, video
analytics significantly improved forensic viewing capabilities.
Today’s cameras, however, trend toward being sleeker, more covert-
type models. Many end users choose cameras that are more aesthetically
pleasing, which doesn’t do much for deterrence. Therefore,
it’s important for end users to take a more proactive approach to security
with their video. We are now seeing a shift toward edge-based
analytics applications providing the necessary intelligence to enable
greater proactivity. For example, glass-break or audio aggression analytics
can provide operators with the opportunity to intervene in a
situation before it can escalate and become worse.
With the Internet of Things trend we are seeing in the security
industry, other hardware technologies can also be complimentary.
As more and more varied systems and devices are connected and
integrated, systems are able to deliver even greater intelligence that
provide operators, guards, first responders and others with valuable
situational awareness to respond to threats or incidents most
effectively.
The analytics market has experienced ups and downs over the
years. Many applications oversold and underperformed which left
both end users and integrators looking for more. There is a new level
of maturity in the market that has realistic expectations and performance
from these types of analytics. The time is ripe for security professionals
to seek out edge-based video analytics applications—and
begin testing them and becoming comfortable with them. These solutions
provide end users with the ability to customize cameras to
meet their particular requirements, remove bandwidth and storage
limitations, and allow for relatively quick adaptation when video requirements
change.
Additionally, customers can take advantage of applications that
help them increase the return on their video surveillance investment
while taking a more proactive approach to security. Remember that
an IP camera is essentially an intelligent network device that happens
to have a lens, so with all the benefits analytics
applications deliver, it just makes sense to harness
that processing power to offload video processing
from the server to the edge.
This article originally appeared in the October 2018 issue of Security Today.