Excerpts from the article featured in Imaging Notes, Fall 2009, pages 27–32.
Matt Bower of BAE Systems discusses the benefits of using the new SOCET GXP® Video Analysis tool to analyze video feeds from unmanned aerial systems.
When Somali pirates hijacked the U.S. freighter Maersk Alabama and took Capt. Richard Phillips hostage in April, a U.S. Navy ScanEagle unmanned aerial vehicle (UAV) built by Boeing’s Insitu unit took video footage of the developing situation. Predator and Reaper UAVs have been in the news recently because the U.S. military has used them to launch missile strikes, such as the one that reportedly killed Osama Bin Laden’s son, against Al-Qaeda and Taliban targets in Afghanistan and Pakistan.
Less publicized is the extensive use of UAVs by the U.S. military for “dull, dirty, or dangerous” surveillance tasks for which they are better suited than piloted aircraft. Real-time images and videos are increasingly used for remote surveillance, intelligence gathering, situational awareness, and decision-making. “At the receiving end, the amount of data available is increasing exponentially,” says Kevin Kelleher, airborne integration lead for the National System for Geospatial Intelligence (NSG) at the National Geospatial-Intelligence Agency (NGA). By associating geospatial information with imagery intelligence, this airborne video surveillance (AVS) technology allows decision makers to view developing situations in their geographic context, track and visualize events as they unfold, and predict possible outcomes.
What UAVs are and what they do
Also known as remotely piloted aircraft, unmanned aerial systems, and simply drones, these remotely piloted fixed- and rotary-wing aircraft and lighter-than-air and near-space systems are both armed and unarmed. They range in size from hand-launched models that look like toy planes and can weigh as little as 12 pounds to the jet-powered RQ-4 Global Hawk, built by Northrop Grumman Aerospace Systems, which has a 3,000-mile range and operates at about 60,000 feet, and the MQ-1 Predator and MQ-9 Reaper, both built by General Atomics Aeronautical Systems.
The Predator, which can fly at altitudes of up to 25,000 feet, performs surveillance and reconnaissance missions and carries two laser-guided anti-tank Hellfire missiles; it can stay in the air for about 40 hours. The Reaper is a larger and more capable aircraft that can fly at 50,000 feet, carrying up to 14 Hellfire missiles, and using infrared sensors to distinguish the heat signatures of rocket launchers, anti-aircraft guns, and other firepower on the ground.
Today, the U.S. military deploys more than 5,000 UAVs, and daily UAV missions in Iraq and Afghanistan have nearly tripled in the past two years. Military applications include peering over hills or buildings, monitoring the seas, eavesdropping from high altitudes, and assisting in special operations. The U.S. military also uses UAVs to transmit live video from Iraq, Afghanistan, and Pakistan. Traditional roles for military airborne geo-intelligence, Kelleher says, include operational support, battle damage assessment, treaty/inspection monitoring, non-combatant evacuation operations, forensic analysis, and coalition operations; new roles include disaster relief, counter-terrorism/narcotics, and homeland defense.
Direct connection between UAVs and operators on the ground or on aircraft is limited to line-of-sight communication; however, communication relay nodes and satellites enable operators to control UAVs and download data from anywhere.
Video analysis software
One of the most valuable products that UAVs can provide is streaming video, in real time or near real time. This product also poses one of the greatest challenges. “To be useful, this massive amount of data must be analyzed in near real time,” says Matt Bower of BAE Systems, a subject-matter expert for the company’s SOCET GXP® Video Analysis software tool. The very flexible and dynamic re-tasking of UAVs, he points out, requires analysts to quickly analyze the data stream, which can be part of live operations.
SOCET GXP, he explains, can take in video from UAVs—plus a stream of support data that includes such variables as the platform’s location, its look angle, the temperature, and the wind speed—and display the footprint of the camera’s field of vision on a map. The user can then move the frames of interest into SOCET GXP to create annotations, mark-ups, briefing products, terrain extraction, building extraction, and so on. “You now have all the SOCET GXP functionality in the video,” says Bower, “and can push the data through. You can fuse the video with other reference sources you might have and use it to drive other geospatial software, such as Google Earth.”
The latest version of SOCET GXP enables users to track moving objects they select on screen, as well as to push the video’s telemetery data into a sensor model and use it to extract the coordinates of the moving object and monitor its speed and heading.
These data streams raise concerns about processing power, says Bower. “If you have a video at 30 frames per second, any advanced computation on those frames—even something as simple as sharpening and dynamic range adjustment—could incur a very big CPU processing cost, because you have to re-do the operation for every new frame.” UAVs can employ various methods to reduce the amount of bandwidth needed to transmit video streams by several orders of magnitude. First, their on-board computers can disseminate only the most pertinent data. Second, they can recognize targets and transmit their coordinates rather than large imagery files. Finally, when they do need to transmit large volumes of data, they can use advanced data compression to reduce bandwidth requirements. Read the complete article >>