Video Detection: Increasing the Light at the End of the Tunnel

Road incidents cause more than 50% of traffic jams – which cause congestion and crashes, according to video traffic incident detection company Citilog vice president Eric Toffin. With more than one billion cars on the road today, reducing accidents, and keeping the traffic flow moving is a challenge for operators and technology is increasingly becoming the solution. Video detection allows operators to identify problems quickly so they are then managed more effectively.

Citilog’s video-based automatic incident detection (AID) solution for tunnels, MediaTunnel, provides automatic incident detection and traffic data in real time through video image processing.

Video incident detection systems have been aggressively deployed throughout the world and in Europe especially since a new tunnel safety directive was released by the EU in 2004. Several major disasters, such as the fire in Mont-Blanc’s tunnel that claimed 39 lives, prompted a reassessment of tunnel safety and a call for video-based AID solutions.

MediaTunnel and Baltimore Harbor, US

With dramatic tunnel incidents an ever-present threat, video-based AID systems for tunnels are increasingly being designed to help save lives and protect infrastructure.

"Accidents will induce congestion and if vehicles don’t stop they will slow down," says Citilog’s Toffin. "Tunnels will very quickly back up or congestion will be created in cities or strategic tunnels such as Baltimore. Our aim is to reduce incident-induced congestion."

About 100 CCTV cameras are monitored by the company’s MediaTunnel system in Baltimore’s harbour crossings network, operated by the Maryland Transportation Authority. The 1.4 mile-long Baltimore Harbor tunnel handles nearly 65,000 vehicles a day across its four lanes and the Fort McHenry tunnel – the world’s widest underwater-vehicular tunnel – serves more than 115,000 vehicles and is eight lanes wide.

With this volume of traffic to deal with it is essential that traffic has a continuous flow and that incidents are dealt with quickly.

"Studies show that every minute you save reacting saves between five to ten minutes of congestion. So if you save five minutes of reaction time you reduce congestion by half an hour to 45 minutes," explains Toffin. "Half an hour of congestion in places like Baltimore has huge consequences. Reducing the reaction time also decreases delays and pollution and so on."

Money matters

"With a traditional system the average response time is anywhere from one to five to ten minutes," explains Toffin. "By using AID the reaction time is reduced to between ten and twenty seconds. When an agency is trying to procure our system, this is what is put forward to the management. AID solution detects incidents more quickly and every minute that you save is going to be valuable."

Toffin believes there is a straightforward economic argument for deploying video detection systems like AID. "We see more and more enquiries from toll authorities wanting to reduce congestion," confirms Toffin, referring to The Millau viaduct in France – one famous example where video detection is crucial.

The Millau Viaduct is a large cable-stayed road-bridge that spans the valley of the River Tarn near Millau in southern France. It is the tallest vehicular bridge in the world, with one mast's summit at 343 metres (1,125ft) – slightly taller than the Eiffel Tower.

"People pay the toll to avoid congestion," says Toffin. "However, if there is an incident on the toll then some will revert back to using the free alternate routes. Every minute of congestion costs hundreds of dollars."

User-friendly software

Last month Citilog released version six of its AID software. It promises fewer false alarms and offers video clips of every detected incident in CCTV quality. However, it is the new interface that will ensure that vehicle detection is made easier than before.

"The graphic user interface has been totally reworked," explains Toffin. "We’ve had the same interface for the last four years and after accumulating wish lists from clients they have been incorporated into the system making it easier to work on."

This has led to version six which includes better performance of algorithms which are video enhanced to allow vehicle tracking at any time with trajectory-based stopped vehicle detection.

With many successful tunnel deployments still being rolled out worldwide – on the French Riviera, China’s Chongqing province, and Paris’s west ring road (the A86) – Citilog’s success in tunnel surveillance makes up just part of Citilog’s video-detection future.

"The future for Citilog and surveillance is in tunnels, but it is expanding," says Toffin.

The deployment of video-detection in tunnels has meant that less sensitive infrastructure has also benefited from the products.

"Baltimore deployed close to 100 cameras and in the next step deployed the system to its outdoor cameras," says Toffin. This type of move is typical by authorities and has proved over time that video detection products like AID reduce risk and economic impact in less sensitive areas.

"Tunnels are viewed as the most sensitive infrastructure so that is where surveillance started, second is bridges – as they are also very sensitive – then highways and expressways. We are seeing an increased activity in bridges and open roadways so this is becoming the trend more and more. We will not move away from tunnels, but will also move to less sensitive infrastructure."