Algorithm Is Gonna Get You: The Advantages of Video Analytics
Here is a real-world assessment of where video analytics are today and the multitude of advantages, possibilities and opportunities they encompass.
Video analytics have gone mainstream, and with good reason. These computer algorithms can provide a better understanding of what’s happening in locations being monitored by video cameras.
They’re quickly becoming a standard feature in surveillance devices and solutions, and have matured enough to offer more practical and reliable functionality than ever before.
SSI consulted with several industry experts to get their perspective on video analytics — and the advantages, possibilities and opportunities that they present. Consider: the video analytics market is projected to grow from $1.7 billion in 2016 to $4.2 billion by 2021, according to research firm MarketsandMarkets. It is an ultra-attractive arena for systems integrators to get acquainted with.
Jennifer Hackenburg, senior product manager with Dahua, agrees: “Protecting a business keeps getting simpler with video analytics and is a great way to differentiate as an integrator.”
Deploying video analytics provides the user with an auditable source of complete data, along with increasing accuracy and consistency in solving surveillance challenges, Hackenburg adds.
“Post-forensic analytics offer a complete picture of what was occurring at a specific area and time of day for conclusive information that is irrefutable,” Hackenburg says. “With this functionality on a system, it’s easier for a business to protect themselves from false liability claims.”
New developments in video analytics are offering enterprise security executives new ways to apply surveillance technology beyond traditional security, making it a new and invaluable asset for business intelligence. According to Hackenburg, the result is that security management can better substantiate the costs to upgrade systems by proving that these systems deliver real ROI when applied across the enterprise.
She points to five video analytics functions built into today’s leading surveillance cameras that not only improve overall security and situational awareness, but can be adapted to deliver a valuable source of business intelligence. They are:
Facial recognition provides the unique ability to identify individuals in real time by cross referencing distinct characteristics against a number of private and public databases and/or watch lists of known individuals.
In addition to its inherent ability to identify wanted criminals, a perfect example is the use of facial recognition to detect cheaters entering casino properties. On the flip side, the same technology can be used to detect high rollers who warrant VIP attention and in retail, it can identify high profile shoppers entering a department store.
Heat mapping measures and rates motion in an image and recolors that image to reflect the intensity of motion in an area. The result is solid data that helps management make effective business decisions. For example, in a retail store, heat mapping can be used to understand the flow of traffic (ex: when people enter the front, do they more often go straight, right, or left?).
Heat mapping also can be used to pinpoint merchandising areas for optimal product placement and provide analysis on why customers dwell in some areas of a retail establishment longer than others. The data provided can offer store management information to determine high volume shopping days and times. Overall, it allows a retailer to make calculated decisions on product placements and the allocation of personnel.
People counting literally counts the number of people entering or exiting a specified area. It’s increasingly being used in sports and entertainment venues, to better manage staffing for concession operations, ticketing areas, and entry and exit points. It’s all based on real-time data, so if a large number of people quickly congregate at an entrance, security personnel can be immediately alerted to make on-the-spot adjustments.
Occupancy estimation can determine on average how many people are present, even at different times of day. Whether it’s a retail store, outdoor concert venue, office building, or sports stadium, the data can help event planners optimize the use of available space, help with staff planning, determine business hours, and even adjust lighting and temperature conditions. In addition, if there is an emergency, security can use the data to communicate with first responders to determine the most appropriate response.
Tripwire/intrusion detection serves as a virtual fence which can detect when intruders are attempting to climb over or ram a physical fence at airports and other facilities that have large perimeters. The technology can easily and quickly detect if an object or a person crosses a virtual line, such as a boundary wall, and then notify security personnel. The right system can alert security operators in real time to suspicious activity, allowing them to verify the alert and then allocate resources and respond immediately.
In addition to its detection capabilities and ability to reduce false alarms (and the costs associated with them), digital fencing can be applied to a multitude of warehousing and manufacturing applications to notify personnel when people or materials cross designated work zones areas, which can help maintain compliance.
Motion detection is an important part of a video surveillance program. It allows security personnel to designate a zone and monitor if people move into or get too close to that area. With this application, security can set up a zone of any shape and size. Any motion in the selected area will provoke an alert and every movement outside of the area will be ignored.
In a museum, for example, the technology could alert security personnel if people are getting too close to an exhibit, or if an object is picked up or moved.
Robert Wegner, vice president system application engineering, Hanwha Techwin, points out that, “In the past, video analytics were a very luxurious and expensive part of a video system that needed large processors. What’s changed is that VA have moved to the camera itself as they’ve become smarter and more powerful. Now it’s become more cost effective to put video analytics on the edge or into the camera.”
According to Florian Matusek, product group director-video analytics, Genetec, artificial intelligence (AI) can help deliver increased accuracy in detection and classification of objects in various ways. But, in the end, it is a tool like others to achieve desired video analytics results.
“Edge analytics can provide added benefit by not requiring server hardware on site; a lot of basic functions can be done on the edge. A disadvantage is that analytics functions are tied to the camera hardware, making it difficult to upgrade the system with newer and better analytics functions,” Matusek explains. “When using servers-based analytics, server hardware can be upgraded much more easily.”
As for predictive analytics, Matusek notes it is still in development and the industry will need to see a lot more results before being able to judge how well it works.
San Kim, director of operations of Phoenix-based Virtual Management Intelligence believes video analytics provide integrators with many possibilities and opportunities to expand their own services, solutions and revenues.
“Video surveillance managed service providers are the driving force that’s directly contributed to the development of better video analytics,” he says. “Integrators, video monitoring companies and in-house security operators would all benefit from video analytics because of its ability to drill down on the specifics in data.”
The technology reduces the time it takes to find what you’re looking for, he contends, enables operators to be notified of specific events and can be used as an early identifier. Among the perks: recurring monthly revenue (RMR) for system integrators and video monitoring companies. It’s also a time and cost saver for both integrators and end users, Kim says.
System integrators are already selling surveillance systems that include video management software (VMS) and cameras. By selling video analytics, they can upsell the surveillance system, providing themselves with additional revenue while significantly increasing the value and return on investment in the surveillance system, notes Zvika Ashani, CTO, Agent Vi.
“For example, if a customer deploys dozens of cameras along a perimeter, adding video analytics for perimeter detection will enable them to monitor that perimeter 24/7 — something that would be difficult for humans to monitor with the same degree of effectiveness,” he says.
Another example Ashani points to would be leveraging video analytics for a software as a service (SaaS) offering, which enables companies providing remote guarding services to receive accurate, automatic, real-time detections of intrusions at remote sites with immediate video verification. By expanding their perimeter protection offering and without having to increase manpower, remote guarding companies can generate a monthly revenue stream with a pay-per-use billing model.
Chris Johnson, regional marketing manager for Bosch Security and Safety Systems, notes that video analytics helps integrators transition away from the traditional one-time sale.
It used to be when installing security contractors sold a video surveillance solution, they sold cameras and hardware and maybe followed up once a year for support and maintenance. Then it was on to the next job unless that customer wanted to expand or upgrade the system.
“With analytics, there are other things integrators can do; they can take the info gathered using analytics and create custom dashboards for their customers to address their specific needs and purposes,” Johnson explains. “Integrators are evolving and can now become more of a consulting partner and solutions provider for their customers. Getting into the business intelligence realm is the next logical step.”
Johnson continued, “From a hardware standpoint, a lot of people can hang cameras but fewer can help customers solve their challenges using analytics. And, it’s an ongoing service they can offer through a cloud based system for RMR.”
For most customers, Johnson suggests about 90% of all the video is never utilized or reviewed. “They record it on the off chance that, if something happens, they can go back to the video and find it. They’re spending a lot of money to only use video say 10% of the time.”
So then the question becomes, he says, is what else can the video be used for? “If I’m going to capture it anyway, and can turn it into actionable information, then I didn’t waste that investment,” Johnson says.
In a retail environment, for example, video that’s initially captured to prevent theft or liability can now also be used to determine more accurately when the store is the busiest, and in which departments. “If the store manager notices that customers typically cut down aisle three, that’s where they’ll put their big promotional displays. Now captured video can be used as a business intelligence tool,” he explains.
Hackenburg concurs. “In addition to security monitoring, analytics can capture relevant business data. Infrared cameras can track customer hotspots, surveillance data can be integrated with point-of-sale [POS] systems to observe sales trends, and data collection allows for predictive analysis that can prevent threats and alert personnel to developing risks before they even occur. Business analytics built into security technology offer incredible growth opportunities for the future of any organization.”
So how can integrators new to video analytics scale the learning curve needed to work successfully in this arena? And, what can or should they expect from their manufacturer partners in the way of support to help them succeed?
Kim notes that “With video analytics there really isn’t a learning curve. The technology is designed to be ready out of the box,” he says. “It has been simplified to a point where you just check boxes of what you want the algorithm to detect, and draw the area for detection. Manufacturers should listen very closely to systems integrators, as they’re the ones that directly interface with the end users and all their challenges.”
Alex Walthers, a business development manager with Axis Communications, advises, “There are so many new opportunities for revenue streams that integrators need to jump into.”
Traditionally, says Walthers, demand is driven by end users and that puts integrators in a reactive mode, meaning they have to learn something new for that account. “I say, ‘embrace the change’ because this will continue. Video analytics are not going away. Integrators should be able to rely on their manufacturer as their No. 1 source of information to get up and running with a new system. Integrators are going to need to get more involved in this, as we see this change coming to the industry.”
Wegner reassures that as long as integrators understand the limitations of analytics and the camera, and how to situate a camera to work as intended, they will be fine.
“Sometimes people have too many expectations. They want to detect something that’s so tiny you can hardly see it on the screen. Everyone has to be on the same page, i.e., know that it only works in these conditions, or in this lighting,” he states.
It’s the same with video analytics, Wegner continues. The technology works well in good lighting conditions, and as long as the installer knows what they’re doing, it is all good.
Hackenburg raises a good point when she says that security management continues to be under intense pressure from the C-suite to justify how their actions, activities and the money they spend on physical security systems add value to the enterprise. It has indeed been a longstanding challenge to calculate the ROI of physical security systems beyond being viewed as a means of providing safety and security and the intangibles of risk mitigation and crime prevention.
But things are changing, she believes, as video surveillance solutions continue to evolve at a rapid pace driven by new advancements in imaging technology and advanced analytics.
From stadiums, to retail, to large enterprises, to critical infrastructure, and healthcare, using data generated from video surveillance systems is rapidly becoming a source of valuable business intelligence. As a result, video surveillance is being used in more strategic ways, and providing evidence of its expanding role and value as a business tool on the enterprise level.
The primary goals of enterprise security are to mitigate threats, to minimize potential losses and liabilities, to maintain a strong brand image, and to keep employees, assets and property safe, Hackenburg says.
“Video analytics continue to evolve as a powerful surveillance tool for enterprise security to detect and mitigate potential threats — often before events occur,” she explains. “Overall, video analytics make it easier and more efficient for security professionals to detect impending and real-time incidents to better protect their facilities.”
Erin Harrington has 20+ years of editorial, marketing and PR experience within the security industry. Contact her at firstname.lastname@example.org.