It's pretty common for a commander planning a mission in Afghanistan to conduct video surveillance beforehand using an unmanned aerial vehicle. But now it's also possible to search through an extensive archive of past video surveillance.
Videos dating back a week, a month, even a year show what has changed at key points of interest such as roads, buildings and terrain, highlighting potential dangers.
"It's kind of like YouTube for the military," said David Barton, who invented the capability with his brother Jason.
All the commander needs, Barton said, is a standard desktop computer, free Java software and adLib software produced by the Bartons' Suffolk, Va., firm, EchoStorm.
It may be as easy to use as YouTube, but adLib is a lot more sophisticated. The stored videos are indexed and searchable and may be embedded with other useful intelligence.
Thus, they can be used to track particular targets of interest, such as a suspicious truck that might appear in different surveillance videos over the course of several days, Barton said.
And by searching through videos by longitude and latitude, it is possible to see how a location has changed over time, he said.
The software and archived videos are being used for mission planning and damage assessments in Iraq and Afghanistan, and to support Predator UAV missions in Pakistan, Barton said.
AdLib is among dozens of technologies, including several similar video intelligence, surveillance and reconnaissance (ISR) systems, being examined by the U.S. Joint Forces Command for their ability to quickly deliver intelligence to troops in combat.
In the adLib system, archived ISR videos are labeled with "metadata tags." These are small strings of computer code that are automatically attached to the videos to provide information, such as longitude and latitude, date, time, altitude, that make it possible for computers to search and find the right videos, Barton said.
Additional tags can be attached manually to serve as bookmarks that point viewers to relevant images within a video, such as the suspicious truck or unusual activity outside a targeted building on a given date.
So, "if you wanted to know what happened at a particular location on Dec. 3 at 10 a.m., you need only enter those search criteria to immediately retrieve the video footage," Barton said.
Some of the searchable video is stored in "federated" databases, meaning that they are linked electronically to create a single large archive whose videos are available to anyone with adLib software and permission to use the archive. As with the Internet, it doesn't matter where the videos are physically stored; the system will find and retrieve them, Barton said.
In other instances, as with the U.S. Air Force's 3rd Special Operations Squadron at Cannon Air Force Base, N.M., intelligence videos are kept in isolated databases and are available only to a select audience.
AdLib is used by the Army, Air Force, special operations units, the Joint Forces Command and the Department of Homeland Security, and at any given time, it is providing videos simultaneously to 200 to 300 users in war zones, Barton said.
But EchoStorm isn't alone in the military "video and data management" arena.
Among the players is the Digital Results Group of Cambridge, Mass., which produces a system called Ageon ISR. The company says its system overlays videos with such intelligence information as photographs, maps and troop locations to create a fuller picture for planning operations.
Ageon, too, promises to deliver data readily to troops in the field. The system, say its developers, delivers "operations and intelligence information to virtually anyone, anywhere using nothing more than a common laptop or mobile device."
On the one hand, the U.S. military wants that. On the other hand, it's not so sure.
"We're very focused on getting ISR data to the tactical edge. That could be a staff sergeant in a Humvee with a laptop," said Air Force Col. George Krakie, the ISR chief at the Joint Forces Command's intelligence directorate.
But after several years of "patting ourselves on the back" for delivering a flood of data to the battlefield, military intelligence experts are having second thoughts, he said.
"Just dumping raw data on the war fighter at the tactical edge does him no good. He doesn't have time to sift through it," Krakie said. What the staff sergeant in the Humvee really needs is intelligence that has been analyzed, he said.
But delivering that is a challenge. The ability of UAVs, cameras and other sensors to collect intelligence is prodigious, but the ability to analyze it is lagging behind.
Consider what persistent surveillance by a UAV yields: "a kilometer-by-kilometer footprint of an urban environment," Krakie said. "A human cannot look at all that information and pull out what's important."
More of that must be automated. "We need algorithms" that will point out possibly important activity so that analysts can then examine it in detail, Krakie said.
The burgeoning capability to deliver intelligence data to troops on the front line has sparked a debate over a practice known as "post before process," Barton said.
Essentially, raw intelligence data is being made available - posted - before it has been processed by analysts.
One philosophy is to "let the end users decide whether it's useful or not," Barton said. "If you're on the ground in the Army going out on patrol, wouldn't you want to see a recent video" of the area about to be patrolled?
The countervailing philosophy is Krakie's - that there is so much raw intelligence that it inundates without illuminating.
One answer is to develop a more centralized intelligence collection and storage infrastructure that has greater capability to analyze data and exerts greater control over their release.
But Barton is betting on letting the user decide. "Post before process is going to be the biggest shift [in intelligence culture] over the next couple of years," he said. ■
E-mail: bmatthews@defensenews.com.
The Army's Future Combat Systems (FCS) program is ready to test a few components that soldiers may have in their hands by 2010.
Help
