Wednesday, June 25, 2008

Brave New World of Video Surveillance

Digital video surveillance is beginning to shake up three market sectors where dumb analog cameras have been mounted for years: public spaces like airports and airbases, colleges and schools, and retail – which include banks and those quick service restaurants and stores.

We spoke with research director Stan Schatt at ABI Research about his recent study “
Watch this Space” highlighting the coming boom in digital video surveillance software and equipment.

______________________________

Why is ABI looking into video surveillance?

We believe in connectivity. The focus at ABI is everything wireless. Surveillance video is rapidly moving to wireless technologies as are we all, so that is why it is interest to us. It is an emerging technology that will fundamentally change global consumer and business markets.

Surveillance video isn’t exactly new.

Video surveillance systems have existed for many years. But until recently, extracting useful information from them was labor-intensive, time-consuming and tedious.

Now, the quickening transition from analog to digital video has made it possible to use software for detection and analysis. This can free humans from the drudgery of slamming through hours of tape, while improving accuracy and the ways video can be use – ways that were just not possible before.

I am particularly excited by the potential of digital video for analyzing and improving business processes. Once the retail sector sees the potential for business analytics, budgets for video surveillance, which now mostly come from the IT department, will begin to flow from the much more generous pockets of Marketing.

Video surveillance software was the particular focus of your study?

Yes. Because we can effectively massage digital records, the potential for intelligent use of software is enormous. Already the link between current physical security applications and what wireless, digital video and its movie-like rich images can do is dramatic.

There are many small software companies in this market, and some big ones such as IBM, which has released software that is largely platform-agnostic, increasing pressure for others to follow suit. While most systems today are sold to end-users, IBM Global Services sees potential in a managed service model, and it would not be surprising to see HP jump in as well, particularly following its EDS acquisition.

There are a number of smaller players who impress. Object Video, which was founded by some DARPA types, stands out with 800,000 licenses.

Some other small companies are quickly becoming familiar names: Axis Communications on the IP-based camera side. PMSC of South Carolina on large storage.

How about airports?

Airports are an obvious application, but you can set the rules with this stuff if you have digital images and digital storage and let it do much of the work. If a bag is left unattended for a number of minutes, the software can quickly, even automatically pick that fact up and set off an alert. If a passenger in an airport corridor suddenly reverses direction or makes an abrupt approach to an exit when the standard flow of traffic across is in the opposite direction, towards an entrance of gateway, there can be an instant alert and intervention.

Smart software can also trigger mechanical consequences – for example, instantly locking down escape routes in an airport when an individual is behaving erratically or takes off at a run.

Any human monitors needed in that digital future?

There’s always room for human judgment. You will see more large institutions – corporations, college campuses – setting up much more elaborate video monitoring centers that go beyond alarms and observation, allowing a shift leader or incident commander to react to what they see happening … or about to happen. Banks already do that.

The metadata procedurals the software designers are playing around with – there are about 13 different detection directions, by my last count – include any number of different characteristics.

What are the most demanding of these?

The most ambitious are automated face recognition programs that banks can use, say, at an ATM – not just to identify bad guys as criminals but to identify you so no one but you can use your card. Smart cameras can also survey the parking lot and tip the central control room to any break-pattern activity, such as leaving a car too long in a particular space … or moving it to a space where it doesn’t belong.

You said in an earlier interview that it reminded you of the futuristic film, Minority Report.

Yes. All this, of course, raises profound questions of civil liberties – and they play with those issues in the film Minority Report, which has Tom Cruise intercepting criminals before they commit their crimes. We have a pretty conservative court system at the moment, so I don’t think you will find much support there in blocking the initial deployment of these software recognition and tracking systems. And who knows what classified smart video applications are already in use in the skies above Iraq that will be coming our way?

What about China? Have the Olympics created any technology leaps there?

I expect there to be some advances, using the Olympics as an excuse for upgrades, but nothing dramatic. The China market is attractive, but a continuing concern, particularly on software, is the potential in that nation for software hacking which is right down doing major damage to the mobile phone market.

Most of the video surveillance equipment in China right now is cheap analog stuff. On the other hand, they have no shortage of people in China who can roll the analog tape back and forth to find what they are looking for. Britain, which began aggressively putting up outside analog video monitoring as far back as the IRA bombings twenty or more years ago, is now aggressively going digital. Most of the British equipment is installed and monitored by local municipalities.

Do you expect industry consolidation in the next year or so?

I do. The video surveillance area in general is so busy with new companies and small firms that I do anticipate some mergers and acquisitions. There are any number of products and platforms in the market, but very little interoperability yet.

The camera companies are also trying to get together, but more by establishing common technology standards than on merging. Canon, Samsung, and Panasonic – not sure why there are all Japanese – are in some sort of dialog on tech standards. North American camera manufacturers tend to focus on the high end. Lumenara from Ottawa is one of those with an 11 megapixel camera coming out soon. You could shoot a respectable motion picture or TV show on a camera that advanced. It’s broadcast quality.

A video stream that rich must take up a good deal of storage.

It does. But once all that information in storage, you can do some very smart things with it. Intelligent processing can begin right inside the camera itself. For example, the digital signal processing chips that are essential to advanced video are faster than ever. They are now able to put some fundamental video analytics into the chip itself. An Israeli firm, Mango DSP, seems to be making some remarkable progress to that end.

Where’s the quickest growth?

The steepest growth curves we see are in the government, retail, and educational sectors. Growth in the government sector is still driven by heightened security concerns since 9/11, of course. We are now seeing video installations in transport systems that keep an eye not only on external surroundings of the vehicles, but on the people inside – including trains and buses.

There’s a popular trend to jump on board a bus after it crashes and then claim whiplash – so come cities are trying to add video to fight that. We’ve all seen the popularity of police video cameras that capture not only police responses, but the provocations that precede the response. The ports are another area where very little of consequence has been accomplished.

How about education and retail?

The video market in the education sector has gone from a trot to a gallop since the campus killings at Virginia Tech – and elementary schools are also now adding video when they can find the budget to do so.

Retail will take the lead, however. It just beginning to appreciate the possibilities of business analytics in all this. Retailers have traditionally low margins, so they are reluctant to take on equipment costs. But if some of this smarter software, backed up with smart chips, can tell them how customers respond to a display and what they like, it justifies the cost to elevate video surveillance equipment and systems – and reduce stock shrinkage by the employees and customers who are stealing them blind.

Tom Goff