updated 9/21/2012 3:15:24 PM ET 2012-09-21T19:15:24

When you think "security" at a political convention, especially one involving the President, you might imagine guards and secret service agents. But at both party conventions this year, some of the most sophisticated technologies to date was and is being used to protect attendees, including President Barack Obama, from threats.

One of those technologies is covert and invisible. It's software being used by local police, the FBI and the Secret Service to comb the Internet for mentions of the President, the convention and any related threatening terms. So-called data mining technology can be programmed to search for specific words and is much faster at poring over websites than any human. What's more, the software uses concepts rather than straight keywords, and its language processing is sophisticated enough to understand context.

10 Trickiest Spy Gadgets Ever

"It sees the word 'bomb' and knows that there's a difference between 'bomb' as in failure, or 'the bomb' as in good," said Bill Bickford, vice president of operations at Jacksonville, Fla.-based IxReveal, a company that produces the data mining software uReveal. The Charlotte-Mecklenburg Police Department started using the software ahead of the convention.

Bickford told Discovery News that the software searches for spikes in activity around specific concepts. When it sees something that looks important, it will flag it so a person can investigate further and decide if it's a threat. A self-described terrorist in South America writing Facebook rants may not be as important as a buzz of activity on a militia site in North Carolina, where the convention is being held. UReveal is designed to distinguish between both.

An added benefit, Bickford says, is that it doesn't need to directly identify anyone. There's no need to store a million data points based on an individual's activity because the software is analyzing behavior en masse.

But that doesn't mean there isn't tech being used to detect an individual's activity. Image analysis software called AISight, available from Houston-based BRS Labs, scans a scene recorded by a camera and flags anything unusual. What it defines as unusual is based on what the software determines is normal. That baseline is established after a camera has been trained on a particular scene over several days.

"So it will say this area looks like this most of the time and this other area looks like that," said David Gerulski, vice president of marketing. When that area of the image changes it will send an alert that something has changed.

"A guy standing at the edge of a parking lot for 30 minutes would get flagged," he said.

In one sense, it does exactly what police are supposed to do -- look for something unusual.

Although the technology was deployed in Tampa for the Republican National Convention and has been implemented in several cities, BRS said they can't confirm or deny whether it's being used in Charlotte.

Hackers Claim Romney's Taxes Ransomed For $1M

If technologies don't warn officials in time about suspicious activity, there is tech designed to track down events when something does go wrong. The Charlotte-Mecklenburg Police Department is using ShotSpotter technology, which locates the source of gunshots if any are fired. The city of Charlotte approved the system in July.

Designed by Mountain View, Calif.-based SST Inc. the system consists of sensors that hear the sound of a gunshot, and by measuring how intense it is, pinpoint its location by triangulating. The company claims a range of up to 2,000 yards for the sensors.

It hasn't been deployed widely in the city -- just near the convention, according to reports. But it will allow police to know where a shot came from and track down the shooter more easily. The city of Seattle was looking at it earlier this year, and the system has been operating in New Haven, Conn.

So far, both conventions have gone on without incident. And that's the way organizers hope it stays, with the help of technology.

© 2012 Discovery Channel


Discussion comments


Most active discussions

  1. votes comments
  2. votes comments
  3. votes comments
  4. votes comments