|
|
 |
|
Law Enforcement
|
 |
|
 |
 |
Too much information
London police find suspects in overwhelming CCTV footage
Nov 1, 2005
|
|
by James Careless
|
If the July 7 London bombings shook the world, the police's subsequent success in identifying the four suicide bombers astounded it. Small wonder: With its deft analysis of CCTV surveillance footage, London's Metropolitan Police not only put faces to the terrorists, but compiled a visual history of their actions prior to the attacks. Suspects in other alleged terrorist bombing attempts from July 21 have been arrested with the help of CCTV surveillance footage.
Clearly, these scenarios were resounding successes for CCTV surveillance advocates. However, it also highlighted some serious issues concerning the timely collection, analysis, and cross-referencing of CCTV footage from multiple sources. To be precise, the omnipresent cameras that allowed the Metropolitan Police to conduct its research also made the task incredibly difficult to execute.
Cameras Cameras Everywhere
If the title of "World's Most CCTV Monitored City" were ever to be awarded, London would be a shoo-in. "There's about half a million video surveillance cameras deployed throughout the city, when you take into account both public and private video surveillance systems," said Peter Fry, director of the CCTV User Group, an industry organization that develops CCTV standards and best practices. "These include cameras in public places, the Underground and on buses, train stations, shopping malls, and banks."
 |
| London police were able to use CCTV surveillance footage to gather information about four suicide bombers from July 7, as well as other alleged terrorists and their accomplices who attempted to detonate bombs in London on July 21. |
Experience shows that London's surveillance cameras deter premeditated crimes such as thefts, but not spontaneous acts such as bar fights, Fry noted. They also help monitor the city for terror attacks, but only make a difference if the terrorists act in ways that attract attention on camera.
Unfortunately, the four suicide bombers who attacked on July 7 didn't stand out, as surveillance images retrieved from the London Underground's CCTV system prove. Hence the city's hundreds of thousands of cameras couldn't prevent the London bombings that killed more than 50 people, but they were instrumental in finding out who did it and how.
With suspects in the July 21 attempts now on trial, the Metropolitan Police is tight-lipped about CCTV's role in investigating the attacks. Still, information from Fry and Avid Technology's forensic video analyst Grant Fredericks make it possible to piece together the process.
After the bombings took place, the Metropolitan Police moved quickly to seize surveillance video wherever it could. This meant dispatching dozens of officers to train and bus stations, banks, private corporations, and any other facilities equipped with CCTV surveillance equipment.
Formats And Tape Stacks
The process netted an estimated 20,000 tapes and digital recordings, said Fry -- a wealth of video equivalent to "something like 200,000 hours of viewing." Where tapes were seized, the process was simple: Officers just took them in hand and left. However, obtaining digital recordings was another matter. In many instances, the officers had to seize the actual hard disk recorders, since downloading the files onto other media would have been "a massive undertaking."
Once the Metropolitan Police started reviewing the footage, "they found themselves dealing with all kinds of different formats," Fredericks told Government Video. "Owning to the lack of standardization and differing ages of CCTV systems, the video was on everything from analog VHS, S-VHS, and digital videotapes to all kinds of digital data formats. It was a nightmare."
As soon as surveillance footage began to come into New Scotland Yard, law enforcement officials began the arduous task of reviewing it. However, both Fry and Fredericks believe that the police are still reviewing surveillance video as you read this, simply because there is so much footage that has to be examined.
"Back in 1994, I had to look through tapes of a riot," Grant Fredericks recalled. "I only had 100 tapes to review of very good quality -- it still took me four months to complete. With the amount of footage available in the London bombings case, they could be reviewing it for years to come."
It's not just the mind-numbing task of reviewing surveillance video that poses a challenge; it is also doing so in ways that minimize having to review the surveillance video repeatedly once potential suspects have been identified.
To avoid this, police video analysts have methodically logged the appearance, approximate age, clothing, and other unique features of the people they see on the footage, entering this information into a searchable database as the video is transferred into a forensic computer system. When a suspect profile is determined, it can be entered into the database's search engine, allowing the relevant footage to be quickly found and called up on screen.
"Say the suspect profile was someone of Middle Eastern origin between the ages of 15 to 30 carrying a backpack," said Fredericks. "By logging in each person's basic characteristics when the video was first viewed, investigators can search the database for people who match this suspect's description." Granted, doing it this way requires video to be located on the forensic computer system, but it is exponentially more efficient than watching the surveillance footage all over again.
No Easy Solution
The sheer bulk of surveillance footage to be reviewed, combined with the manpower required to view it and enter notes into the database, has turned the investigation into a titanic undertaking. So what about technological shortcuts? There must be ways, for example, to reduce the analytical workload using biometric analysis, right?
Unfortunately, the answer is no. There is no quick fix available. A case in point: Although biometric analysis -- mapping the faces of suspects mathematically, then comparing these maps against the faces captured on surveillance video -- has been undergoing field tests for years, it's still not up to the job.
"At this point in time CCTV doesn't provide enough image resolution to do biometric analysis reliably using current technology," Fredericks explained. "As well, you need a controlled environment in which to capture images for biometric analysis, in order to provide images capable of comparison. Once HDTV finds its way into CCTV, this might change, but right now biometrics is not up to the job."
Obviously, the sheer volume of tapes and digital files that had to be reviewed during the investigation has been enormous. Yet, despite this, the department did manage to spot the bombers relatively quickly, allowing them to track down alleged accomplices in a matter of weeks.
"Overall, the Metropolitan Police did a fantastic job," said Fry. "And they did it in an incredibly short period of time, just a matter of days."
So how did the Metropolitan Police make it happen? First, they analyzed the most relevant footage first, such as video recorded on the subways and buses. Second, they already knew where CCTV systems had been installed throughout the city. This made it possible to retrieve the footage quickly, with a minimum of confusion. (Fry will reveal more details during his talk at the GSPX conference in Santa Clara, CA, later this month.)
Lessons Learned
Ironically, the challenges faced by the Metropolitan Police might actually allay fears of futuristic Big Brother observation, because when half a million cameras are watching a city, there are just too many feeds for any one agency to monitor in real time. To event attempt it would require astronomical manpower costs.
Also, the belief that CCTV feeds could be converted to IP and then shipped digitally to a central monitoring point is also unfounded. "The network doesn't exist that could carry all this traffic without crashing," Fry said.
With the ever-present possibility of terrorist attacks in North America, what can we learn from London's experience in cracking this case? Fry noted that CCTV surveillance for public transportation (and, for that matter, public areas in general) is a necessity for the deterrence and investigation of crime. The successful use of CCTV by London police should serve as a main argument for promoting the idea that more systems are needed.
Law enforcement agencies also need to create a comprehensive database of CCTV systems in their area. The FBI developed the model for such an initiative during its pre-event security operations for the Winter Olympics in Salt Lake City (see Government Video, May 2002). Agents compiled a database of civilian surveillance cameras in the area so they could quickly retrieve potential evidence in case of an incident.
While widespread CCTV surveillance resulting in a police state is still fantasy, the reality is that its presence in London made it possible for police to identify the July 7 bombers. Without CCTV, these bombers might not have been identified, making the chances of finding their accomplices virtually nonexistent.
MORE INFO
Avid Technology www.avid.com
CCTV User Group www.cctvusersgroup.com
GSPX conference www.gspx.com
|
|
|
Government Video is part of the NewBay Media Community
copyright 2007© NewBay Media LLC.
810 Seventh Avenue, 27th Floor New York, NY 10019 T (212)378-0400 F (212)378-2160
|
|
|
