The January 2011 issue of SSI features “Video Continues to Enhance Its Security Image,” in which a dozen top technology providers were asked to identify the latest and greatest video surveillance technologies industry professionals can expect to have the highest level of impact during the year, and of course video analytics was high on the list.
Cernium Corp. CEO Craig Chambers took some time out to talk with SSI about the advances of video analytics, and why he believes the technology will be featured in virtually every DVR and most IP cameras on the market. So what’s new with video analytics in 2011? Read on as Chambers tackles that question and explains how the technology can aid integrators in growing their business.
Can you explain video analytics?
Craig Chambers: Fundamentally, video analytics is a form of machine vision or computer vision that evolved from being able to look at or pick out objects and still images to the ability to do it with video that is rapidly moving images going through the machine. It’s evolved incredibly in the last 10 to 15 years.
When we talk about analytics, the capability can vary all over the map. Some manufacturers will suggest that they’ve got video analytics if they simply have motion detection built into their machines. When we talk about video analytics, we’re talking about a much higher level of capability. That’s really the ability to discern objects and to some extent classify them by asking if it’s a real object. If it is, is it a person, a vehicle or some other kind of object? Then you have to catalog the way that it moves through the scene and use all of that information to determine whether it is an important piece of information that needs to either be stored or forwarded to someone.
As we jump into 2011, what is the latest and greatest in video analytics?
Chambers: I think the latest and greatest is that it’s gone from being a technology that everyone was talking about to a capability that’s pretty much built in from one form or another to most high end cameras, DVRs and other surveillance equipment.
For the first time, most folks in the industry that are dealing with any kind of high end video technology are embedding that analytical capability into their devices as a standard feature. So while that may not be hugely exciting to say, it’s usually exciting to the industry because we’ve gone beyond just offering it as some curiosity that only a few people use to one that is pretty much expected as a feature in those high end devices.
I think the thing that is most new and different is the ability of these devices to be set up very easily. It used to be joked that you’d have to send a Ph.D. along with every system 10 years ago in order to configure them, adjust them and make them operate. That’s changed radically in the last couple of years, and we’d like to think that Cernium is one of the companies that has made that possible. If you look at the user interfaces that we create — and a handful of other companies are attempting to create as well — we reduce that configuration burden to a simple point and click that pretty much anyone who can operate a computer and understands basically what a video system is supposed to do can operate it. Very intuitive interfaces are probably the new announcement for 2011.
What are the greatest challenges that video analytics faces?
Chambers: As a technology, the challenges are not as big as they once were. However, there still is a pressing question that needs to be answered: How do we get something that requires a lot of computing power into devices that pretty much anyone can afford? I would say that there are a large number of folks in the industry who still haven’t figured out how to answer that question yet.
General providers still haven’t quite figured out how to do the really high end analytics in lower cost and affordable devices that the whole security industry can take advantage of. At lower cost, they’ll tend to give you quite a bit less capability and probably lean towards what I would describe as motion detection. That’s really not video analytics. The challenge is getting the price points to a level where all of that functionality can really be utilized in regular, everyday devices.
What is being done right now to address those challenges?
Chambers: There are a handful of companies — like ours — in the industry that have figured out how to write software to a level of sophistication that allows that functionality to take place with a lot less computing power. As a company, you’ve either learned how to do it at this point or you haven’t.
In terms of where it’s going, I suggest that within just a year or two, you’ll see it in pretty much every DVR and most IP cameras that are on the market. Other than that, I think the great story here is about the diversification of capabilities that have now been created using that basic technology. The technology has really gotten a lot more sophisticated and diverse now, so it’s everything from that line crossing, and in some cases facial recognition, which is sometimes considered to be an analytic technology, to object classification. The great leap forward has been figuring out how to integrate that information the same way you’d integrate other types of data into regular user interfaces that security system operators would typically use for their jobs. The neat stuff is what you don’t see, which is all the sophistication that was necessary to make it fit in with the regular video management system or security system operator user interfaces.
Why should installing integrators care about video analytics? What kinds of opportunities will they get from both services and market niche standpoints?
Chambers: In terms of being able to provide capabilities that couldn’t be provided before to their customers, everyone should be interested. For instance, one of the great user benefits is if you use it properly, you can radically reduce the amount of storage that is necessary to keep relevant video. [This means] that the same DVR that might have only held a couple of weeks’ worth of video can now store months and months of relevant video with all of the unimportant information filtered out. That same filtering capability, which is really fundamentally what analytics is doing, can also be used to filter and direct information from video management recording systems to people wherever they happen to be — across the Web — throughout any kind of network system.
So instead of having people manually go through a lot of video or watch video all the time to see if something interesting is going on, which is at best a horrendous waste of time and at worst a very expensive undertaking, you can now use the analytics technology to filter and then direct only relevant information to people through the network to their mobile phones, to PCs to wherever they happen to be at network endpoints and not waste their time and be able to act very quickly on the relevant information as it is required.