Mystery solved - The unwatched video of CCTV systems the world over
15 frames per second. Full HD image. 24 hours per day. 30 days of recording. It adds up to around 900 Gigabytes of footage stored on disk, and all for a single camera. When you expand this out to say a small office environment with perhaps a dozen cameras and it's already over 10 Terabytes, before even considering the likes of university campuses, transport networks and critical national infrastructure.
A lot of footage, that's the bottom line. And with the new trend of "video off" monitoring, with screens left blank until the system's programmed analytics pick up an anomaly, what potential answers to mysteries are lurking on this dark-data that never sees the light of day?
Reuniting absent-minded commuters with misplaced belongings? In amongst the lunchboxes, sweaty gym kits and shopping bags there are surely incidents of lost sentimental items, possibly from departed loved ones that can leave the individual in a state of real anguish. While some transport networks might utilise video for cataloguing their lost property, what about those things dropped in the streets?
Where company cameras include public areas these kind of incidents could be captured and used to inform people of the fate of their property.. Or perhaps identify instances of fly-tipping in rural areas where private cameras cover more areas than those controlled by local government.
More intriguing perhaps are the potential butterfly-effect incidents lurking amongst lots of disparate CCTV systems; events that are seemingly innocuous if viewed in isolation, but when combined with others paint a much more sinister picture..
TV shows like Spooks, 24 and Fauda all show counter-terrorism departments using surveillance footage to identify their target's activities, but even these fictitious representations of government-backed surveillance highlight where tracking is not possible due to insufficient surveillance coverage.
Privately owned and operated surveillance cameras vastly outweigh those under government control and if these smaller systems were combined in an effort to trace individuals, who knows how many interactions with potential suspects could be recorded and chased up. When working on identifying "lone wolf" attackers anything the authorities can glean from their day to day activities could be the key to finding the architect before they slip back underground.
Then there are the potentially criminal acts that fly under the radar, instances of corporate sabotage or espionage; the vagrant rummaging in the bins behind the office may not be cause for concern, but if the same figure is seen posting things through a competitor's front door suddenly there could be more to it than meets the eye.
Of course the notion of sharing recorded footage for on-going police investigations is not uncommon and warrants to review private camera systems are routinely issued. Similarly, the freedom of information act allows individuals to request CCTV footage of themselves on a specific time/date/place basis. Even the method of bringing these disparate recording archives together has been - at least in-part - addressed, but these tend to only be compatible with the more enterprise level systems leaving the local fish and chip shop unlikely to be integrated.
The problem is any implementation of such a system, be it local, regional, national or even global is that of security and privacy. An "Overwatch" system may seem like a step further towards a big brother state of ultra surveillance because in many ways, it is!
Who would have ultimate control? One answer would be to deploy distributed ledger technology to ensure that there were no single authority with ultimate control, but then what of those joining the Overwatch? Potential misuse abound and how can legitimate intent be gleaned when the system's owners are thousands of miles away, behind a set of IP addresses?
Well, this too could be mitigated if rather than providing "eyes on" access to the users of the Overwatch, the machines are given control. After all the analysis of frames of video is much better suited to an algorithm anyway. Feed it a set of parameters and then matching instances could be anonymised and perhaps access to the video could be again, anonymously requested and the video only being released when agreed by both parties?
Who knows if this kind of collaboration will ever see the light of day in the more sensitive, guarded and paranoid world of today, but what if it held the key to real transparency?
Given surveillance of the entire world is only going to become more complete maybe time is better spent working on how to openly share the data for good than opposing the train that's already left the station.
