Movement for the masses

In the wake of the atrocious 7/7 suicide bombings in London, Mayor Ken Livingstone had this to say: “We are increasing CCTV coverage. There will hardly be a park or a transport system where we haven’t installed surveillance systems.” A statement obviously made to reassure and encourage citizens to continue their use of our Capital’s public transport networks.

Nonetheless, and hardly surprisingly, passenger throughput on the London Underground dropped by 30% across August last year.

Somewhere in the pecking order that lists cost and reliability, a major priority that the public desires from any transport system is to feel safe and secure when using it. If they don’t, they’ll not be a regular customer. It is here that surveillance technology can have a significant impact, both directly and indirectly.

At the most basic level, surveillance can reduce the presence of graffiti, loitering and general anti-social behaviour and, at the highest level, help to prevent incidents pertaining to national security. While it would be somewhat naïve to suggest that modern surveillance techniques and technology could put a definitive stop to terrorist attacks, they are nonetheless rapidly becoming an invaluable player in the overall security infrastructure.

Ken Livingstone’s comments were clearly well intended. Certainly, the presence of more CCTV cameras – combined with an increased police presence – will provide a level of reassurance and act as a deterrent. That said, any increase in sheer numbers of surveillance cameras does engender several discussion points.

The surveillance challenge

Mass transportation has several unique characteristics. It has a large number of distributed locations, high-density people movement, multiple vehicles (trains, buses, trams, ships and aeroplanes) and an abundance of CCTV cameras. In turn, these characteristics create two major challenges for the CCTV specialist, the first of which is solely concerned with real-time monitoring and event detection.

The ability of surveillance teams to effectively monitor and gain security insights in real-time is often limited due to information overload. Typically, at any given time only a small proportion of monitored locations can be seen in a Control Room, severely curtailing the probability of ‘dubious’ activity being detected by way of human observation. Not only that, human perception is such that small changes are often ignored. While training undoubtedly helps, lots of busy images can distract operators from smaller – though potentially far more significant – changes.

The consequence is that events of major significance might easily be missed. Even if an incident is noticed some time later, it’s time-consuming and difficult to find information that enables the situation to be assessed such that the decisive action needed may be taken. Confused and/or delayed responses can have serious consequences for security. Improving information availability while avoiding information overload is therefore a priority. That will enable fast decision-making, reduce response times and increase incident detection rates.

Next generation ‘smart’ security technology solutions are now offering end users a genuine breakthrough in terms of their capability. Advanced digital video and content analytics – that includes motion and abnormal event detection, vehicle and unattended baggage detection – is now enabling law enforcement and transport surveillance teams to automate what would otherwise be rather arduous and tedious detection tasks, prevent and respond to threats in real-time and effectively investigate and reconstruct criminal and security cases.

These ‘smart’ event management tools enhance awareness by alerting the surveillance team when a suspicious event is in progress. They provide all of the relevant information and guidance while also keeping track of the current live situation.

While it would be somewhat naïve to suggest that modern surveillance techniques and technology could put a definitive stop to terrorist attacks, they are nonetheless rapidly becoming an invaluable player in the overall security infrastructure

Airport security teams are among those providing the vision. One such is operational at the Changi Airport in Singapore. Changi is served by 74 airlines, connecting Singapore to 170 cities across 53 countries. The airport is the second largest in the Asia Pacific Region and is recognised worldwide for its efficiency, facilities and passenger security. When Terminal 3 opens later on this year, it will expand the airport’s capacity by 20 million passengers to a massive 64 million.

To address the requirements of significantly more complex passenger security needs, the Changi Airport management team has decided to enhance its ability to detect and identify potential security threats. Advanced video content analytics integrated with its CCTV cameras will boost airport security.

The project means that Changi becomes the first major international airport to launch the deployment of video content analytics on such a large scale, providing the on-site surveillance operatives with the power to identify risks in real-time, make assessments and take action that improves security efficiency.

Real-time threat detection

An excellent example of such automated real-time threat detection centres on the identification of unattended luggage. As soon as the CCTV system recognises that a bag has been left unattended, an alert is immediately raised which then enables the surveillance team to easily retrace the steps of suspicious individuals. It can be quickly verified whether or not the event in question constitutes a security risk, and the appropriate course of action may then be taken.

The end result? Improved control of potential security threats, the prevention of unnecessary terminal evacuations and resultant cost savings to tenant airlines. On top of that, passengers may be spared frustrating delays.

The second challenge that these new systems address is in incident investigation. To find the proverbial needle in a video haystack when a security event occurs is a considerable challenge. It can be very difficult to pinpoint the relevant people or objects out of the huge amount of unstructured information captured by hundreds – or even thousands – of cameras. Investigation requires a good deal of time and dedicated resources, as the police have demonstrated in the wake of 7/7 and the attempted bottle bombings involving Heathrow Airport in recent weeks.

In essence, there are four steps – or ‘filters’ – that can be applied to video to help highlight and prioritise digitally recorded content and help reduce investigation time, increase process effectiveness and enable the better distribution and use of resources. These are as follows:

(1) The identification of video footage that includes the movement of objects such as people, cars and trains, and thus eliminates ‘irrelevant’ video. These filters use a specially-designed, high performance video motion detection mechanism that’s particularly sensitive to direction and region of movement.

(2) Automatically displaying suspicious or abnormal behaviour (for example, people running, jumping, bending down or forming crowds) without having to pre-define every possible abnormal behaviour.

(3) Mark out pre-defined susceptible locations and situations such as intrusion into train tunnels, tampering with the rails, vehicles blocking critical access areas, loitering or bags left unattended on platforms. These filters perform with low-to-medium traffic density, and enable better resource use (for example, using security personnel for high density traffic areas and filters for the other areas).

For both real-time event detection and incident investigation, a centralised security and surveillance operation is absolutely vital. In the future, security and surveillance systems will need to be viewed in a more holistic manner

(4) Automatically create a ‘photo album’ of peoples’ faces as they pass a camera, thus reducing the investigation time and helping with any subsequent forensic activities. Typical use is to capture people photographs while they are using escalators and gates. An advanced Best Practice is to employ mega-pixel high resolution cameras that will capture high quality static pictures while video is ‘caught’ at CCTV quality.

Centralise your operations

For both real-time event detection and incident investigation, a centralised security and surveillance operation is absolutely vital. In the future, security and surveillance systems will need to be viewed more holistically. Making sure video is transported back to the ‘nerve centre’ is key.

A number of trains and buses in the UK are now equipped with CCTV and digital recording capabilities. The recorder is usually stored on the bus. As a deterrent – and for evidential purposes – this may prove crucial for minor offences, but as part of an integrated surveillance operation its usage is somewhat limited.

Other metropolitan mass transit systems administered by numerous rail and bus operators have a need to improve security in and around stations as well as protect strategic infrastructure locations (among them bridges and tunnels). Some excellent examples exist Stateside.

As the largest State-wide public transportation system in the US, New Jersey Transit (NJT) provides in excess of 750,000 daily trips for commuters and those on excursions, operating 240 separate bus routes, two light rail lines and 11 commuter rail lines. It is responsible for the efficient and safe transportation of millions of commuters. The operation uses digital video and audio recording for its state-of-the-art Secaucus Transfer Station, recording from more than 200 cameras integrated with 16 DVRs for first-rate security throughout the passenger terminal.

Other advanced analysis tools can add valuable data to the video database (such as the number of people on a concourse, or the registration numbers of vehicles passing by a checkpoint).

Toll road operators, meanwhile, are looking for systems that will increase customer throughput and prevent gridlock when paying to enter a toll section, streamline applications such as the detection of disabled or stranded vehicles and unauthorised parking, and provide booths with protection from random acts of vandalism.

Making surveillance smarter

All of these surveillance solutions are operational in airport and transport systems across the world, and deployed by forward-thinking security teams that have recognised the problems inherent in simply adding more cameras. Making surveillance smarter, using digital video networking, recording and analysis maximises information availability and reduces data overload. It equips the security team to face new challenges that require faster, better-informed decision-making.

More cameras may indeed be necessary in central London and in cities like Birmingham and Manchester, but on their own may prove insufficient to deliver the security improvements the public, passengers, transport operators and the security services require.