The London Eyes
- By John Williamson
- Nov 09, 2007
The United Kingdom has become CCTV nation. Conservative estimates place about 10 percent of the world’s CCTV cameras in the U.K. The consensus is that the U.K. has more cameras per capita than any other country.
Another widely quoted statistic is that U.K. city residents are captured on camera up to 300 times per day on average. Penny Hayward, spokesperson for the CameraWatch CCTV industry advisory body, says nobody is quite sure of the real tally because nobody has actually counted all the installations. In fact, reckons Pierre Hagendorf, chief technology officer of unified visual communications specialist Radvision, there may be as many as 5 million CCTV cameras in the country, or one for every 12 residents. Other observers put the total even higher than that.
The U.K.’s world leadership in deployment of surveillance systems puts London high in the running for global capital of the video surveillance business. The city’s networks are operated by the Metropolitan Police Service (MPS), the Transport for London (TfL) authority, the Department of Transport and local government councils. This is in addition to many thousands of private CCTV set-ups operated by schools, sports stadiums, parking lot owners, shops and banks.
What’s certainly not in doubt is the fact that the UK’s capital city itself literally bristles with CCTV. Events ranging from the dramatic— such as the U.K.’s largest cash robbery in London’s commuter belt in 2006 and the failed nightclub bomb attacks this year in the city’s West End -- to the more workaday -- such as the on-going enforcement of the city’s congestion charging scheme and the monitoring of bus traffic lane regulation -- are all captured on video cameras operated by, and sometimes shared between, a growing number of different operators.
The MPS surveillance network is a major element of London’s CCTV capability. This system is said to manage some 30,000 camera feeds, although the MPS did not respond to interview requests regarding this story. As well as the MPS’s own surveillance points, the capital’s police force has had access for some years to the CCTV systems set up by local government authorities and the TfL organization, which monitors road traffic, buses and trains, stations and transportation hubs. Until quite recently the local authorities would provide video footage when the MPS requested it to help with information about previous incidents. However, with improved technology, the MPS is now able to work more closely with each local authority to share live images in real time and communicate with officers on or near the scene.
MPS Overhauls C3I
The large-scale integration of the city’s various video surveillance networks is a key component of a larger MPS initiative to overhaul its command, control, communications and information management (C3i) systems.The entire program involves several different projects and includes the integration of 32 borough control rooms into a Central Communications Command (CCC) housed in three purpose-built special operations centers in the city.
In August 2006, Tyco Fire & Integrated Solutions–Traffic & Transportation won a multi-million dollar order to supply the MPS CCC with what was then thought to be the world’s largest video monitoring and control system for public safety and security.
Although Tyco declined to address specifics, its system is believed to include a custom video matrix with in excess of a quarter of a million crosspoints acting as a virtual switching matrix, and uses the TV Network Protocol (TVNP) to communicate between switching nodes. TVNP, according to Tyco, will enable further integration of borough and stadium CCTV systems with virtually unlimited expansion in the future.
“The C3i system allows all the cameras that are connected to the system to be viewed and controlled from the Met Police control centers, so in effect they are completely interworked,” observes Neal Entwistle, marketing director of Tyco Fire & Integrated Solutions - Traffic & Transportation.
“The significance of TVNP is that it is an open protocol standard that allows many disparate systems to work together and is the backbone of how the systems are connected in London for the C3i project.” Two of the three CCC operations centers are up and running, with the third due to go live before the end of the year. Lambeth, one of the two sites already in operation, according to Shushmul Maheshwari, CEO of RNCOS, a market research company, will eventually handle more than 500 public events each year in the capital. RNCOS has published several studies of the CCTV industry, the most recent being “Global CCTV Market Analysis (2007-2010).”
Despite Entwistle’s description, others are not so quick to declare TVNP is an open standard in the true sense of the term.
“The jury is still out,” says Barry Keepence, chief technology officer of IndigoVision, Edinburgh, Scotland, which makes end-to-end IP video and alarm management security solutions for CCTV surveillance applications. “TVNP is not widely deployed.”
Neither has it been certified by any standards body, Keepence adds, and IndigoVision has installed a number of large surveillance systems that don’t use it.
What's In Store?
The government’s July decision to exempt TfL and the MPS from certain provisions of the U.K.’s Data Protection Act 1998 sparked the increased collaboration between the MPS and other video network operators. The government decision permitted the bulk transfer of automatic number plate recognition (ANPR) data -- used to police London’s congestion charging system -- from TfL to the MPS. In a statement to Parliament, Tony McNulty, Minister of State, Home Department, said: “The MPS requires bulk ANPR data from TfL’s camera network in London specifically for terrorism intelligence purposes and to prevent and investigate such offenses. The infrastructure will allow the real-time flow of data between TfL and the MPS.”
The announcement of the MPS ANPR deal caused something of a stir with unconfirmed reports that the MPS was looking to store ANPR traffic data for five years, and that managing and storing what would be the multiple exabytes of data captured in that period could cost as much as $3.5 billion. Regardless of what happens with the ANPR material, data storage could become more of a concern for everyone in the U.K. CCTV industry as the switchover from analog to Internet Protocol (IP) and megapixel cameras gathers pace, says Chris Williams, marketing director of Wavelet Technology.
He calculates that whereas a conventional analog CCTV image might have a file size of 50 kilobytes (KB), a 5 megapixel camera file could go to between 250 and 400 KB.
Given the advantages that the newer technologies can bring to CCTV data analysis,Williams suggests that image compression or recording at a lower image rate might bring only limited relief. “Megapixel cameras are going to enhance the ability to do more forensic work,” he contends. “I don’t think the file sizes are going to get smaller because you need the detail, and the detail implies a larger file size.”
In parallel with pressure on file size and storage, as more U.K. video surveillance systems become networked and distributed, despite the bandwidth efficiencies of IP transmission bandwidth may become more of an issue for users, especially those who rely on leased lines or other carrier arrangements.
Asked what he considers to be the challenges attached to managing, switching and monitoring a CCTV system with thou sands of feeds, Hagendorf responds:“Bandwidth. Bandwidth. Bandwidth.”
The fact that the IP video industry as yet lacks a comprehensive body of standards has implications for bandwidth. For sure there is widespread use of the MPEG4 standard, but as Keepence of IndigoVision notes, “MPEG4 standardized the decoder. How good an encoder is... is left up to the manufacturer. How good the compression is depends on the implementation of the encoder.”
As a consequence, claims Keepence, the data rate for some CCTV solutions to give good quality video can be in the range of 2 to 10 megabits per second, while those from IndigoVision cut in at around 1 Mb/s.
A further consequence of this lack of robust standardization is that it’s not simple to mix and match different CCTV components from different vendors.“Although standards such as MPEG4 exist, there are differing interpretations of these and it is not generally possible to, say, use one manufacturer’s IP encoder and then decode the stream with another manufacturer’s device,” points out Tyco’s Entwistle.
But Keepence advises not to blame interoperability problems on poor standards definitions. Using cable TV set-top boxes as an analogy, he points out that MPEG2 signals used in the cable connection will work with any manufacturer’s cable box.The settop remote control, however, will not work with another manufacturer’s cable box.
That it doesn’t has nothing to with the MPEG2 standard, he says.
And here it’s maybe also worth noting there’s some residual reluctance to make a full commitment to IP video systems due to perceptions about their possible vulnerability.
“There is, however, still a resistance in the U.K. for major infrastructure projects to switch to pure IP surveillance because of the perceived security risks associated with doing so,” acknowledges Steve Gorski, managing director,Axis Communications (U.K.) Limited. “As such the networking of surveillance systems still involves Axis and its channel partners in a great deal of education as to the benefits of network video and the real business continuity and risk management-linked benefits offered through centralised monitoring, management and storage of video output.”
Depending on the chosen architecture, network bandwidth may also be a concern for those operators introducing video analytics, a technology where, in Maheshwari’s judgment, the U.K. is at the vanguard. At present most video analytics system deployments in the U.K. -- and in many other locations too for that matter -- are still in the trial stage. One such pilot is currently running at London’s Clapham Junction station, by some measures the busiest rail station in the country.The primary goal of this trial, which uses a system supplied by Agent Video Intelligence Inc. (Agent Vi), Ft. Myers, Fla., is to gauge the effectiveness of video analytics in graffiti detection.
In general video analytics uses image processing algorithms and other technologies to automatically detect and alert operators to out-of-the-ordinary events and occurrences. According to Agent Vi’s founder Gadi Talmon, video analytics started out with a firm security focus, but down the road is finding applications in traffic management, retail, and education, as well as graffiti detection.
There’s even the prospect of boosting revenue in using analytics in retailing, for example, to automatically track and count customers and identify ‘hot’ and ‘cold’ spots in a store.“What is the effectiveness of promotion and so on,” comments Talmon. “This is very valuable information.”
Unlike some other video analytics solutions that do their processing in one location -- either in the field or on a server -- Agent Vi systems use a distributed architecture called Image Processing over IP (IPoIP).This has a “lightweight” agent algorithm in the field and the major algorithm on the central server. The company claims several advantages for this arrangement: the server is highly scalable, there are bandwidth efficiencies because only the results of the field analysis are transmitted back to the server, and the agent can be readily embedded into a variety of existing video devices.
Talmon predicts that scalability will become key in video analytics systems, and that more and more IP cameras will be supplied with analytics pre-installed.
There are those who believe that the hoopla surrounding video analytics is somewhat overdone. While accepting that the technology, if properly deployed and configured, can be a very useful tool, IndigoVision’s Keepence suggests: “Analytics is one of the most over-hyped and overpromised and under-deployed technologies in the whole CCTV industry.”
Nevertheless the U.K. Home Office’s Scientific Development Branch is currently engaged in a major effort with industry to standardize what it terms Video Based Detection Systems (VBDSs). Dubbed the Imagery Library for Intelligent Detection Systems, or i-LIDS, this effort is the government’s benchmark for VBDSs and consists of a video test library of CCTV footage designed to evaluate systems for government use. Manufacturers meeting the highest level of performance classification will be entitled to use the trademarked i-LIDS logo in their trade literature. The initial 2007-2008 i-LIDS evaluation schedule is focusing on four monitoring and detection scenarios: Sterile Zone (June 2007), Abandoned Baggage (September 2007), Parked Vehicle (November 2007) and Doorway Surveillance (February 2008).
As they say, watch this space.