Video Analytics: Mature Solution or Still-Developing Technology?

The early struggles of video analytics technologies to live up to over-inflated expectations are well documented. Thankfully, improvements in performance have gone a long way toward overcoming end users’ and integrators’ disappointment and frustration with these solutions. That said, no amount of technological advancement will completely erase the stigma of the “over-promise and under-deliver” days, which is why many integrators still haven’t completely embraced video analytics wholeheartedly.

“We’re definitely moving ahead with caution and finding good analytics partners that can stay with us and make sure we deliver a good product for our customers,” says Brent Edmunds, president of Stone Security, a systems integrator based in Salt Lake City. “We’ve shown a lot and we’ve piloted a lot, but to be honest, we still haven’t deployed a ton.”

While some technologies have matured to the point where integrators are comfortable deploying them, for some security providers, customer demand simply isn’t there yet.

Generating Revenue From Analytics

In addition to improving security — and in many cases overall operations — one of the most attractive benefits of deploying video analytics for security integrators is the potential to generate recurring monthly revenue (RMR). One way Simpson Security earns RMR is through the traditional 24/7 live video monitoring, and the revenue is “considerable from this service, as it is per-camera, per-month,” according to Keith Simpson.

To streamline and improve this service, Simpson Security employs video analytics to reduce the number of alarms and/or notifications cameras generate to eliminate nuisance alarms and free up storage space, which can be costly.

“The simple form of analytics is having the camera ignore a tree blowing in the wind and cars going down the street right outside the protected area,” Simpson says. “Analytics is a plus, and there is a need for it in this type of protection, which enhances the system and its reliability and can be customized for the client depending on their needs.”

For a line-crossing deployment, Allied Security Links (ASL) generates revenue by setting up mobile solar trailers that are monitored by its monitoring partners, who receive an alert whenever someone or something enters an area that is supposed to be unoccupied during certain times of night. For example, this might be used for a construction site that’s active Monday through Saturday from 8 a.m. to 6 p.m.

“On Sundays, alerts will pop up at the monitoring station, and they can view video and either call the client or the police, depending on what they want done for each situation,” says ASL’s Christopher Tamez.

“I haven’t seen a whole bunch of demand for video analytics,” says Tristan Soule, senior sales and design engineer for integration firm A3 Communications, Irmo, S.C. “A lot of what I’ve seen are license plate recognition and people-counting “For some detention centers, we’ve been trying to do a virtual fence so if anything crosses a line, it will set off alarms with the surveillance camera.”

Stone Security may be taking a cautious approach to video analytics, but the company has had success with tools such as Briefcam and Snap, which some may or may not consider to be actual video analytics solutions.

“Today, it’s been more of the tools like Briefcam for forensic review of video or Snap, which stitches together video to provide a storyboard of connected cameras within the system to track and trace people through the system,” Edmunds says. “Those have been more applicable for our customers, who are using Briefcam in some pretty interesting ways; and they’ve been adding quite a few analytics to their platform, such as facial recognition,” he adds.

Real-World Deployments

According to Keith Simpson, president and CEO of Simpson Security Systems in Alexandria, La., the most common way his company’s customers use analytics is for using virtual fence or tripwire applications. Other common uses made possible by advanced camera technologies include flame and smoke detection, wrong direction alarms and people counting.

“Cameras nowadays are motion sensors that detect changes in pixels and can detect an object passing a predetermined point or boundary. They can also detect if an object is human or animal, depending on size and weight,” he says.

Face the Future

If there is one type of video analytics that appears poised for growth, it would be facial recognition, which Brent Edmunds of Stone Security says customers have been asking about more frequently.

“Facial recognition is the one that stands out for me,” he says. “The next step for the security industry is providing really solid, reliable facial recognition or at least understanding when and how to apply it. I’m not so sure people are doing that quite right at this point,” he says.

Tristan Soule of A3 Communications agrees. “I know the VMS guys are doing facial recognition and there are third-party companies doing it, and that’s what seems to be the biggest thing, especially for bigger crowds — sports arenas or very busy lobbies and such,” he says.

When setting up line-crossing locations for triggering recording on the camera, Simpson says his installers also create bookmarks in their VMS of choice, Hanwha Techwin America’s Wisenet Wave, to speed response and post-event investigation.

“This allows for easy retrieval of specific type of events on the software,” he says.

Vision Security Technologies of Birmingham, Ala., mainly deploys video analytics for establishing tripwires along fence lines and for providing after-hours notifications when people are present in a parking garage. The company has also seen some demand for people-counting and retail analytics for identifying hot spots based on foot traffic, checkout line length and other factors, says Barry Komisar, the company’s founder and CEO.

“The retail analytics obviously help with product placement in the store for the end user and improve customer satisfaction by notifying management if checkout lines are long so additional checkout lanes can be opened,” he says.

One of the biggest benefits of video analytics is their ability to support operators engaged in live video monitoring, Simpson says. The longer someone watches video on a monitor, particularly when they are viewing multiple streams from multiple locations, the more likely they are to miss important events captured by any of the monitored cameras. Analytics, however, are immune to this phenomenon.

“Video analytics can improve the effectiveness of live monitoring by offering the ability to detect a missing or added object that either was not there and now appears or that was there and is now missing,” Simpson says. “We’ve deployed this analytic and used this approach in airports, large warehouses, car dealerships and utility security.”

A more specialized use of video analytics Simpson Security has deployed is to use license plate recognition for damage control for valet-parked vehicles. The approach involves multiple cameras and analytics to view an entire vehicle and capture any dents, scratches or other damage to a vehicle before a valet drives it.

“If the vehicle owner states that the establishment damaged their vehicle, they can go back to footage that shows damage was present prior to dropping it off at the valet,” Simpson says.

To AI or Not to AI?

One of the most talked-about technologies in security — if not in general — is artificial intelligence (AI), which offers the promise of making surveillance and other systems more effective by “learning” over time. While the technology may eventually deliver on that promise, from a real-world perspective, it isn’t quite there yet, and — at least at the moment — could end up causing headaches for integrators, believes Tristan Soule of A3 Communications.

“I know everybody thinks AI is going to be the next biggest, best thing in the world but it’s like anything that’s been introduced into this space. It takes like at least five years to get through and then it becomes big,” he says. “With it being new technology, you’re going to have your hiccups; so until it’s perfected, it’s hard to go out there and put something in that you have to service once a week.”

Christopher Tamez, CEO of Allied Security Links (ASL), Austin, Texas, says his firm has also employed license plate recognition, in this case for controlling access to a customer’s location via a front gate.

“We connected the gate to the camera so that when a license plate is recognized that is on the white list, it will open when the client designates,” he says. “For example if they are doing a training class ... and only want the gate to open when a designated vehicle arrives with a license plate that has been entered into the system, it will start opening at 9 a.m. and stop working at 9 p.m. Any time before or after that, they would have to call in; plus, the client will be notified by email and text that an authorized vehicle has approached the gate outside of hours giving permission to be on site.”

In addition, Tamez says, ASL set up a handover feature on the Hanwha PTZ camera that allows it to lock in on and continuously follow an object. And by combining the camera’s edge technology with software provided by Arteco to integrate with the VMS, users benefit from a single platform, which streamlines the process of retrieving information compared to what would be necessary with multiple platforms.

The Verdict

Without question, video analytics technologies have vastly improved in recent years, offering security integrators a variety of tools they can deploy to improve their customers’ security and streamline their operations.

“I’m not somebody who’s going to try to sell something or pitch something to a customer if they won’t really get any value out of it,” Soule says. “I’d rather they spend the money on something they need but don’t have, like another camera to cover a blind spot, than on something they’re probably never going to use.”

The final verdict about whether to deploy video analytics for customers depends solely on determining what, if any, benefit they will get from these solutions.

“Don’t get me wrong: we do sell analytics. It’s just that it’s not a need for everybody,” Edmunds says.