Five ideas for video surveillance
I'm taking the bulk of this week's recap to summarize some interesting ideas I heard this week while attending the MIPS event in San Antonio. Here we go...
Idea #1: The A&E Problem
A&Es and integrators are not working together particularly well. This was the gist of a presentation from video consultant Charlie Pierce, who said that while specifying work for security projects often falls upon the shoulders of A&Es, they often don’t have the skill sets in security they need. One of the challenges Pierce sees is that A&Es have traditionally operated off of manufacturer spec sheets, which historically reflected inaccurate specs (Pierce operated a warranty/repair center for CCTV cameras for years, so he knows the industry products inside and out -- literally). His recommendation was for integrators to develop better relationships with A&Es and that integrators offer up their knowledge to A&Es. What no one wants to see, said Pierce, are A&Es that spec projects which can't realistically be delivered by any integrator.
Idea #2: The RFP Problem
This little jewel came from a Q&A with MIPS speaker Terry Haukom. Haukom spearheads the Minnesota Department of Transportation's roadway camera project. The question was about RFPs that are selected on price alone. That, said Haukom, is the sign of a bad RFP. He suggested that a well-written RFP should be written such that it measures who is selected upon two ingredients: 1) price and 2) value.
Idea #3: The Lens Problem
Cameras have a lot of power, and it's becoming more common to see amazing processing capabilities in them. Vendors are putting multi-megapixel designs on the market, and some are even offering full-frame rates at super high resolutions. At the current rate of development for camera image processing and frame rates, it's not hard to picture a day in the not-so-distant future where surveillance cameras are cranking 30 frames per second at 20 megapixels. So where does that upward spiral of camera power end? It's hard to say where exactly, said Fredrik Nilsson, general manager of Axis Communications. Instead, Nilsson says the limiting factor is going to be whether we can create high-quality camera lenses to match the resolution of future cameras. "The lens is becoming the limiting factor for megapixel cameras," Nilsson said. "The lens is a mechanical element; it doesn't follow Moore's law."
Idea #4: The Edge Explodes
Nilsson said that by the end of 2011, the processor chips inside many IP video cameras would be able to handle 90 percent of all video analytics algorithms. That's a huge "edge" capability. Paul Bodell of camera maker IQinVision added that the "edge" is also being driven by edge storage -- flash card storage within the camera itself. At some point, solution designs might not even need central repositories; there will be so much storage in the camera that centralized storage might only be of concern if your cameras could themselves be vandalized or destroyed by accidents.
Idea #5: The Cloud Explodes
This idea was voiced by Lars Thinggaard, the President/CEO of Milestone Systems. "It is a matter of time before the math ends up in favor of the cloud," he said. What's particularly interesting is that Milestone isn't what you might call a cloud company. They have offered software basically to create local network video recording -- on-site, on a local computer, or even in a server room. However, what you've seen in the company over recent years is development of architectures so that access to this video becomes more cloud-like, and I would expect this trend to continue and affect the entire industry. Thinggaard said the upswell of cloud adoption for video surveillance is likely to come first from the lower end of the market and move upward from there in terms of adoption.
In other news
TSA workers under investigation, Gunmen raid California tech company, more