Are Public CCTV Cameras an Utter Fiasco?

The UK's love-hate relationship with CCTV cameras continues. This time, the head of the UK's police video imaging division called their use an "utter fiasco." Setting aside the prudence of creating an international bruhaha, it's quite damming given the UK's leadership in the use of CCTV technology.

For background information on this debate, see the BBC's report on why CCTV is failing to cut crime, and its report on why CCTV is not being used properly. For general background information, see our special report, "Is Public CCTV Effective?"

What's the Problem?

The UK chief stated, "we've got enough cameras, let's stop now, we don't want any more cameras.'

The problems stated:

• Tedius job of shifting through recordings
• Lack of a system to retrieve, process and distribute CCTV crime scene images
• Too much spent on equipment, too little spent on people and processes

What's the Solution?

The BBC article mentions that, in an effort to improve effectiveness, that the "unit is also looking at ways of using software which can follow distinctive brand logos on the clothing of unidentified suspects." Probably not a good idea to try scientific experiments to solve an utter fiasco.

Spend More on People?

While I am sure that operator training has value, the technological problems are likely to be too severe to overcome through mere training. Systems that cannot interoperate and systems that require extensive manual video review only benefit so much from training.

Implement PSIM and Video Analytics?

There are two clear emerging product categories within video surveillance that address these issues. PSIM would provide a sytem to retrieve, process, distribute, etc. Video analytics would reduce the amount of time sifting through video.

Why the Solution is Hard?

Even if you had the money and the will to implement these solutions, the logistics are daunting.

Retrofitting video analytics on existing cameras can be expensive, complex and hold long term risks. Do you change out the current analog cameras for smart IP cameras? Do you add in intelligent encoders at the camera? Do you add analytic servers at the headend? How do you tightly integrate it with video management systems? Plus, forensic video analytics are not mature (e.g., searching for objects, colors, etc.) and incompatible with one another - creating risks of poor performance, limited use and demands for upgrades in the next few years.

How do you deploy a PSIM solution for the government? Do you choose one PSIM provider and then integrate everything to it? Since there are no standards and hundreds of video management manufacturers, how do you integrate them all together? What happens when some systems do not fully function because of limitations of the recorder?

Most of this is not technologically revolutionary (though tracking brands on clothing would be). In the absence of widespread system interoperability standards, it's just logistically difficult to pull all of this together without spending a billion pounds and not facing risk of failure. The cost would likely be this high even if you only integrated a fraction of the UK's estimated 3-4 Million cameras (assuming a conservative $1,000 USD cost per camera to add in analytics and PSIM).