(CJ Burton)
It had the makings of a public-relations nightmare. Recently a blogger working for a local daily newspaper phoned the management of Rogers Centre. He’d been watching a baseball game on television the night before and noticed in the background a child walking down an aisle carrying a glass of beer. Mario Coutinho, vice-president of stadium operations and security for the facility and its home team, the Toronto Blue Jays, assured the blogger they would never permit such a thing. The newspaper remained unconvinced. “They accused us of selling alcohol to a minor,” Coutinho recalls.
He needed to get to the bottom of the issue, and fast. Reviewing the televised footage, he established during which inning, and in which section of the stadium, the boy had been sighted. That helped him pull the relevant security camera footage. “We were able to go back and find out that his mother was with him,” Coutinho says. “She gave him the beer, she stayed at the top of the stairs and said, ‘Here, go give it to your dad.’ And we watched him go down the stairs with the beer and give it to his father.” Coutinho showed the footage to the blogger, and spared the Rogers Centre some potentially embarrassing news coverage.
Four years ago Coutinho would have had little to show. Back then 32 analog Panasonic cameras watched over the stadium, part of a system dating back to its construction in the late 1980s. It was something of a shambles: over the years cabling and cameras stopped working. Since replacement parts were often unavailable, half the cameras eventually went blind. Recording equipment, meanwhile, failed regularly. Even when the system functioned, however, it was quite incapable of capturing footage in the necessary detail. The cameras were pan-tilt-zoom models: an operator would have to have trained one on the boy at just the right moment. Zoomed out, the boy and his family would have resembled formless blobs. [adspot]
But that was then. In 2010 Rogers (which owns Canadian Business) began laying new fibre-optic cable throughout the stadium. Then it installed a new network of high-resolution security cameras from Vancouver-based Avigilon. The cameras covered all 50,000 seats in the stadium’s bowl, and Avigilon’s control software tied them in with the older analog cameras that still watched hallways and stairwells. Radio and binoculars close at hand, Coutinho routinely monitors baseball games, football games and rock concerts from a small suite perched above the masses, surrounded by LCD screens displaying real-time surveillance footage.
Avigilon is a young company struggling to place itself at the forefront of an ongoing transition from analog to digital surveillance—a revolution that is altering what security cameras are and what they can do. Since its first commercial sales in late 2007, the company has become one of the fastest-growing technology companies in North America and earned a small profit of more than $6 million during the first half of this year. Rogers Centre and other stadiums are among the early adopters of digital surveillance, and with good reason—the larger the area to be monitored, the more sense Avigilon’s ultra-high-resolution cameras make.
This year, for the first time, industry-wide revenues from digital sales are expected to surpass analog in North America and Europe as a wider array of customers migrate to networked surveillance systems. Not only does this transformation signal an opportunity for digital-only Avigilon to grab market share, it also opens the door to using the humble video camera in novel ways that have nothing to do with security—like improving how crowds flow through stores or helping public transit operators keep track of their rolling stock. The problem is that surveillance cameras are still typically purchased using security budgets (which tend to be small) and by security specialists (who often don’t understand how surveillance might be applied to other parts of the business). Can Avigilon founder and CEO Alexander Fernandes find enough customers who are as forward-looking as he is?
Fernandes was uniquely qualified to shake up video surveillance. A veteran of digital imaging technology, in 1999 he founded Quantitative Imaging Corp. (QImaging), a manufacturer of digital cameras for scientific and industrial uses—the kind, for example, that can detect microscopic, early-stage cancer cells. He sold QImaging just three years later for a reported $20 million but remained there two more years before departing. At age 34 he was independently wealthy and ready for a new challenge.
The digital revolution had arrived late to surveillance. Analog cameras typically capture video at VGA resolution—640 pixels across by 480 pixels vertically. That’s fine for monitoring a narrow hallway or small convenience store. Put one in a medium-sized parking lot, though, and it’ll be darned near impossible to make out facial features or read licence plates. The problem is exacerbated because the majority of analog cameras are plugged into recorders that can’t even record at VGA quality. Instead, they compress the incoming signal, resulting in further image degradation.
The abysmal quality of footage is a problem, because at the end of the day, it’s the recordings—not the live surveillance—that’s most useful. A security guard’s command post with a dozen surveillance monitors may appear to afford a superior view of what’s going on inside and around a building. It’s a mirage. One study by the U.S. National Institute of Justice found that “after only 20 minutes of watching and evaluating monitor screens, the attention of most individuals has degenerated to well below acceptable levels.” The footage, after all, is the polar opposite of engrossing. Six hours into an eight-hour shift, one screen could show a safe being dynamited open and a guard likely wouldn’t notice.
For this reason, security cameras are most often used to investigate incidents after the fact—provided the imagery is of sufficient quality to be of probative value. Analog persists primarily because it’s cheap. (IMS Research, a market research firm specializing in electronics, estimates this new generation of surveillance cameras cost on average 4½ times more than analogs.) And unlike digital, the technology is standardized: plug an analog camera into any VCR or TV, and it will work. Installers don’t require special training. The owner of your local convenience store is probably quite happy with his analog camera, thank you very much. [adspot]
Fernandes first became acquainted with video surveillance after some of QImaging’s computers and inventory went missing. The company responded in the usual way, installing better locks, a burglar alarm, pass-card entry and video surveillance. Accustomed to high-end medical and scientific imaging systems, Fernandes was appalled by what he saw. “It occurred to me that the world needed and would want something better,” he recalls.
After leaving QImaging he reflected more about why video surveillance seemed so backward. The explanation, he concluded, lay in the fact that companies active in the sector tended to specialize in cameras or recorders or software, but rarely offered the whole gamut. The resulting lack of standards proved an insurmountable barrier to innovation. One could sell the best digital security camera available, but if there wasn’t video management software available that took advantage of its features, what was the point? And good luck trying to find an installer with sufficient expertise to make various systems from different vendors work together seamlessly. Fernandes concluded that to disrupt the market, he’d have to master both software and hardware to create a full suite of products.
He and fellow QImaging alumnus Wan Jung founded Avigilon in 2004. Starting with a four-man team, they gradually recruited engineers, physics graduates, software designers and mathematicians specializing in algorithms. In this Fernandes was aided by his extensive contacts at universities doing high-end research, which comprised much of QImaging’s customer base. “It’s not even thousands. It’s probably in the realm of hundreds of designers and engineers that do this kind of thing,” he says. But he already knew where to find them—and many of them lived in Vancouver.
Some of what Fernandes wanted to accomplish wasn’t feasible: the necessary processing power and communications bandwidth either didn’t exist or was prohibitively expensive. Fernandes gambled it would be commercially available by the time he needed it. “If you have to wait two years longer than you’d expected for the technology to catch up, you can go bankrupt two or three times over,” he concedes. “But that’s where I’ve made my whole career, playing on that razor’s edge of what’s possible and what’s not possible.” By the time he’d spent his first $30 million on R&D, the needed technology was on the market. Avigilon began selling its first high-resolution camera in 2007.
What truly sets Avigilon’s cameras apart is their eye-popping resolutions. Its first camera boasted 11 megapixels, which Fernandes claims is higher than what any competitor offers today. Its flagship offering is now 29 megapixels—comparable to what you’ll get from high-end digital SLR cameras used by professional photographers. By way of comparison, other companies selling high-definition cameras rarely offer much above two or three megapixels.
Higher resolutions aren’t just about bragging rights: they open up new possibilities. At Rogers Centre, 19 of Avigilon’s 16-megapixel cameras simultaneously cover all 50,000 seats in the bowl, as well as bullpens and suites. To accomplish the same coverage with analog, Coutinho guesses, would require at least 400 cameras. “By the time you look at your cost of cabling, electrical, managing it, it’s almost impossible,” he says.
Avigilon developed a technology called High-Definition Stream Management, which enables its cameras to compress, transmit and record high-quality footage. It also has its own video management software called Avigilon Control Centre. What’s more, Avigilon’s software is designed to make nice with other technologies, including analog. That flexibility is crucial: security systems are often upgraded over time, as budgets allow. Iain Morton is a regional vice-president for Tyco Integrated Security, which sells systems from Avigilon and its competitors. He says this flexibility represents one of Avigilon’s unsung strengths. “They have the least expensive and one of the most effective encoders for taking an old analog camera into their new digital platform,” he says. It’s crucial for customers who want to migrate to digital over time.”
Avigilon’s approach has worked remarkably well thus far. From a standing start, its revenues have more or less doubled every year, surpassing $100 million last year. “International expansion and new product launches are driving above-average industry growth,” observed analysts at National Bank Financial recently. (Analyst praise is almost embarrassingly effusive: all seven covering the stock rate it a Buy.) And the company now employs 400 worldwide. Avigilon’s ambition is to become the world’s largest video surveillance company. By offering what he regards as a superior product, Fernandes thinks that goal is within reach. “The idea is to flood the market and fill the sales channel with Avigilon as fast as possible, before anybody catches up.”
It’s not far-fetched. Video surveillance is a surprisingly fragmented market, with no player controlling more than 5% of the whole. It’s also growing rapidly: IMS Research estimates total industry revenues could reach $29 billion a year by 2016, nearly double that observed today. That’s partly due to strong demand coming from Brazil, Russia, India and China, where investment in developing infrastructure for surveillance of cities and major sporting events has been heavy. And Avigilon has something most competitors don’t: National Bank Financial analyst Kris Thompson observed that whereas most competitors serve only specific customer segments (the most important of which include education, policing, retail, transportation, city surveillance and others), Avigilon’s offerings address most conceivable applications, and it is also building software modules to address specific requirements.
Then again, many of Avigilon’s competitors have greater resources. Axis Communications of Lund, Sweden, the current market leader, has been around since the 1980s and did six times more business than Avigilon last year. Last year Axis dedicated 500 employees and 15% of its revenues to R&D, and has committed to what it calls an “offensive release rate” of new products and services. “They are still a dominant force,” says John Woodhouse, an analyst with IMS Research. Axis is active in North America, reporting strong interest from retailers and the public sector in Canada. Many competitors enjoy widespread brand recognition—Panasonic, Bosch and Samsung, for instance.
In preparing annual reports on the video surveillance market, Woodhouse has seen many moon-shot companies come and go. “You get a lot of these companies such as Avigilon who will get the formula right for a few years and boost their market share,” he says. “There’s a few every year.”
In Greek mythology, Argus was a hundred-eyed giant of surly disposition. You might consider him the god of video surveillance: Somehow, Argus’s eyes relayed useful information to what must have been an impressive occipital lobe that could make sense of it all. Avigilon and its competitors are trying to accomplish something similar in the real world: they want to create systems that can see, interpret and recall everything with ease.
High-resolution cameras represent only a first step. The next step is what the security industry calls “analytics.” An early example is the ability to read licence plates. Once a digital camera captures a licence plate at a parking lot’s entry or exit point, another technology called optical character recognition (OCR) can read it and convert it into text. That information can then be used in any number of ways. A parking garage operator, for example, can integrate its system such that it can alert attendants whenever a vehicle known to leave without paying enters the garage. Or security personnel could be warned when a recently fired employee’s vehicle arrives on company premises.
Future analytics should be more impressive—for example, software that can recognize suspicious objects (like an unattended package) or behaviour (such as a person loitering too long in a given area) and alert security personnel to check things out. Although the necessary technology exists, Fernandes says it’s too expensive and unreliable. “The accuracy levels are down below 80%, and they probably need to get up to 95%–98% for it to be usable,” he says. “If you’re conducting city surveillance, and you get 800 alerts a day, you might as well get 10,000.” [adspot]
Digital technology and analytics could also be applied to things that have little to do with traditional security. A nuclear power generation company, for instance, had Tyco install surveillance equipment in irradiated areas of its plant. The plant’s actual security system was sourced from a separate contractor—these cameras offered a means of conducting inspections, reducing the need to dispatch personnel in protective gear.
The San Diego Metro Transit System installed an Avigilon system to monitor its platforms, yards and parking lots. Although that’s used in many ways you’d expect—such as detecting vandalism—it’s also used to find trolleys. “It sounds funny, but every once in a while one of these trolleys will go to the wrong track, or they’ll lose track of it,” Fernandes explains. “Before Avigilon they’d have to send some guys walking around, sometimes for hours, looking for it.”
Such applications require linking systems that previously didn’t communicate. Many surveillance companies are therefore acquiring technologies outside their traditional area of expertise and integrating them. In May, Avigilon paid $17 million for RedCloud Security, a privately held provider of physical and virtual access control systems, and recently completed integrating its products with Avigilon’s products. That level of integration opens up interesting possibilities. For example, a manufacturing plant could use analytics to facilitate overnight deliveries. Rather than hiring a person to handle graveyard-shift shipping and receiving, the plant could use its system to create a virtual barrier that truck drivers can’t cross without tripping an alarm. If they do, they’ll get an automated warning telling them to go back where they belong, or perhaps check in with remote security personnel on an intercom. The shipper can even wave the signed-off waybill in front of a camera; OCR software can read it and send pertinent information to the customer’s logistics systems.
Such examples hint at how Avigilon’s customers can use its surveillance systems to improve their operations—and raise revenues. Thing is, these systems are still purchased primarily using security budgets—and since security systems seldom generate revenue, those budgets are tight. If Avigilon could persuade customers its systems are actually business tools, the cost of purchasing them might be shared by operations, human resources and other budgets. That’s actually something of a Holy Grail across the entire security industry. But customers willing to entertain such new ideas are few and far between. IMS’s Woodhouse says retail is one obvious industry where algorithms and analytics could be married with security cameras; one could monitor customer traffic throughout stores, for example, in hopes of better managing floor space. But at many retailers he’s familiar with, business intelligence people don’t talk to security personnel. It’s the same thing in casinos, where jazzed-up security systems could track high-interest areas and tables on the floor. “There’s not that many forward-thinking casinos in my experience,” Woodhouse says. “A lot of people don’t recognize the potential yet.
“The growing feeling is, when’s it actually going to happen? Who’s going to be the first to make it really widespread? There’s certainly a few companies that could do that, Avigilon potentially being one of them.” Fernandes will be walking the razor’s edge between possible and impossible a while yet.
