Video and analytics moving further toward the perimeter

Intelligent devices and video surveillance are becoming important components of perimeter security, and could drive growth in that segment
 - 
Wednesday, December 18, 2013

ENGLEWOOD, Colo.—The growing trend of deploying video surveillance and analytics devices for perimeter security is showing no signs of abating. The market is poised to reach $160 million by 2017, up from $90 million in 2013, according to a report from IHS Research. 

IHS expects North America to rank as the largest market for this segment of the perimeter security market, according to Blake Kozak, senior analyst, access control, fire and security, at IHS. Spending on sensors, which will remain the backbone of perimeter security solutions, will continue to outstrip spending on video and intelligent analytics for these applications. 

“While video analytics and cooled and uncooled thermal cameras will continue to gain traction in certain end-user industries, sensors will continue to provide detection for most end users,” he said. “Analytics can be useful to detect beyond a perimeter fence line and track that person once they have breached the perimeter, but sensors will still be required.”

As a component of perimeter security, video surveillance and intelligent analytics fit well within the layered approach that has long been the foundation of outer perimeter security, Kozak said. This philosophy, which integrators typically recommend, is predicated on the idea that multiple types of detection maximize the strengths, and minimize the weaknesses, of the various elements of an integrated perimeter solution.

“By using multiple sensor types, the probability of defeat is lessened while the probability of detection is heightened,” he said. “Additionally, multiple layers can reduce false alarms when integrated.”

Though intelligent analytics have improved since their introduction to the market, false alarms continue to plague end users, Kozak said. One area, however, that shows particular promise in terms of false alarm reduction is behavioral analytics, an evolving field of intelligent software that, unlike rules-based analytics, enables cameras to absorb information and make decisions with little programming required. “Behavioral analytics do not require rules to be assigned or trip wires to be programmed,” Kozak said. “Instead the camera is installed and over the course of a few days, weeks or even months, the software will learn the events occurring in the FOV (field of view) and will learn what is normal and what is not.” 

Houston-based BRS Labs is a leading player in the behavioral analytics field, Kozak noted.  

While technological developments continue to reshape perimeter security solutions, certain high security applications will continue to use guard patrols in conjunction with video surveillance and sensors, Kozak said. He added, however, that video is often deployed to reduce demand for guards, which perform an important function by providing the rapid response. To this end, some end users who deploy video for perimeter security are using it in conjunction with wide-area mass notification, also called giant voice, allowing security operators to view live footage of an intrusion in progress while notifying the intruder via loudspeaker that they are being videotaped and the police are en route.

One of the most talked about trends in the video for perimeter security discussion, according to Kozak, is the “slew-to-cue” functionality. Whether integrated with ground surveillance radar or a PIR, this solution will detect moving objects and automatically move the camera’s view to the location of an alarm, tracking the object in motion as it moves across a camera’s field of vision, Kozak said. Once the object leaves the field of view of one camera, it is then picked up seamlessly by another, and tracking continues. 

“This is a hot topic for the industry because it allows for automated video tracking,” Kozak explained. “Instead of a security operator having to find the incident at a fence line manually with the camera, with slew-to-cue, the camera automatically pans to the alarmed location.”

Market growth could be hindered by “high profile system failures,” Kozak said, such as an instance in 2012 at JFK International Airport, when a stranded jet skier breached the airport’s perimeter fence and crossed two runways before reaching a terminal. The breach prompted the Port Authority to reevaluate its $100 million perimeter detection system, Kozak noted. “Events such as these can damage the successes that the industry has had, which include higher probability of detection and reduced false alarms through more advanced algorithms,” he said. 

Some technological concerns also persist. Bandwidth remains a sticking point for server-based analytics, Kozak said, as does the need for redundancy in the event of power failure or error. He added that servers typically have a limit on the number of channels, and as companies grow they may need to purchase new servers to accommodate expanded security needs. 

Kozak said these ongoing concerns could intensify the debate among suppliers over which solution is preferable: server-based analytics versus embedded devices. While both have their merits and drawbacks, questions linger about whether end users deploying embedded devices may be sacrificing some of the robustness afforded by server-based solutions, which allow for more flexibility in terms of rule creation and analysis, Kozak said. 

Still, loyalties may come down to the needs of a given end user. 

“Many argue that server-based analytics are generally used by more high security end users, where filters can be put in place and more rules can be applied and analyzed,” Kozak said. “Embedded devices are thought to be best for retail and commercial applications."