The Magazine of the National Intelligence Community
Future Technologist Dr. Peter Highnam Director Intelligence Advanced Research Projects Agency
www.GIF-kmi.com
July/August 2013 Volume 11, Issue 5
Video Enhancement Software O Social Media Big Data O Commercial Remote Sensing
Intelligence
KNOW-HOW
When it comes to intelligence, we know how to deliver global solutions that provide a decision advantage and prevent tactical and strategic surprise. Uniquely trained, experienced and qualified, the intelligence professionals at DynCorp International support the strategic, operational and tactical needs of the Intelligence Community. Our trusted intelligence training, analysis and mission-support solutions help our national, military, civilian and law enforcement customers reduce costs and achieve new levels of performance and productivity. DynCorp International provides our customers with unique, tailored intelligence solutions for an ever-changing world.
www.dyn-intl.com
GEOSPATIAL INTELLIGENCE FORUM Features
Cover / Q&A
17
New Horizons for remote sensing
7
As it copes with tightening limits on its military and intelligence markets, the commercial remote sensing industry is reinventing itself by developing new products and finding new ways to manage data. By harrison donnelly
10
21
25
Full motion video technology is becoming an increasingly important tool in military and intelligence operations, and efforts to develop technologies that make that video more useful and valuable to analysts, commanders and warfighters are keeping pace. By Peter Buxbaum
New analytic tools are helping military and intelligence users harvest the rich crop of data available through social media. By Karen E. Thuermer
Data triage can provide the key for time-constrained intelligence extraction from newfound data troves of unknown size, characteristics or content. By Samuel L. Park and Chris A. Mattmann
Video enhancement
Departments
July/August 2013 Volume 11, Issue 5
Social Intelligence
Unlocking big data
Industry Interview
2 Editor’s Perspective 4 Program Notes/PEOPLE 14 Industry Raster 27 Resource Center
Richard m. Cobb
President and Chief Executive Officer TerraGo
28
Dr. peter highnam Director Intelligence Advanced Research Projects Agency
EDITOR’S PERSPECTIVE
Geospatial Intelligence Forum Volume 11, Issue 5 • July/August 2013
The Magazine of the National Intelligence Community Editorial
Managing Editor Harrison Donnelly harrisond@kmimediagroup.com Online Editorial Manager Laura Davis laurad@kmimediagroup.com Copy Editors Sean Carmichael seanc@kmimediagroup.com Laural Hobbes lauralh@kmimediagroup.com Correspondents Peter A. Buxbaum • Cheryl Gerber William Murray • Karen E. Thuermer
Art & Design
Art Director Jennifer Owers jennifero@kmimediagroup.com Senior Graphic Designer Jittima Saiwongnuan jittimas@kmimediagroup.com Graphic Designers Scott Morris scottm@kmimediagroup.com Eden Papineau edenp@kmimediagroup.com Amanda Paquette amandak@kmimediagroup.com Kailey Waring kaileyw@kmimediagroup.com
Advertising
Associate Publisher Scott Parker scottp@kmimediagroup.com
KMI Media Group Publisher Kirk Brown kirkb@kmimediagroup.com Chief Executive Officer Jack Kerrigan jack@kmimediagroup.com Chief Financial Officer Constance Kerrigan connik@kmimediagroup.com Executive Vice President David Leaf davidl@kmimediagroup.com Editor-In-Chief Jeff McKaughan jeffm@kmimediagroup.com Controller Gigi Castro gcastro@kmimediagroup.com Trade Show Coordinator Holly Foster hollyf@kmimediagroup.com Receptionist Vania Jones vaniaj@kmimediagroup.com
With the help of Que, his artificial intelligence personal assistant avatar, new analyst Steve does some career planning and prepares to meet with his supervisor to discuss performance objectives. Senior analyst Sally spots some unexpected changes in a construction project under observation, and asks her assistant, Pixie, to pull together information on a piece of machinery identified in a recent photo. Supervisor Samantha holds a “conversation” with her avatar, a small blue circle named Ed, about the employee appraisal process. These hypothetical intelligence personnel are the possible future benefiHarrison Donnelly ciaries of a recently launched project by the National Geospatial-Intelligence Editor Agency to create the Geospatial Metaverse, an immersive virtual environment for the production and management of integrated intelligence regardless of physical location. First put forward as a concept by science fiction author Neal Stephenson in his 1992 novel Snow Crash, a metaverse is a network of 3-D virtual spaces joined together as a virtual universe, as in one of today’s popular massive multi-player online games. As outlined in a request for information from the NGA InnoVision Directorate, the Geospatial Metaverse would use cost-effective off-the-shelf technology to develop a collaborative, web-based environment blending virtual worlds with geospatial and other intelligence information. The office offered the above “day in the life” narratives to illustrate how the system might work. The goals of the initiative are to “deliver a revolutionary capability to discover, use and return geospatial data for the Department of Defense, intelligence community and other organizations; enable collaboration [and] allow for cooperative confrontation and synergy of analysis; evolve the GEOINT tradecraft to meet 21st century missions; and change traditional concepts of achieving superior organizational performance.” Another key concept of the project is to implement user experience development technology to help employees achieve a “state of flow”—the timeless state of relaxed concentration on a task described by psychologist Mihaly Csikszentmihalyi in his seminal work, “Flow: The Psychology of Optimal Experience.”
Operations, Circulation & Production
Operations Administrator Bob Lesser bobl@kmimediagroup.com Circulation & Marketing Administrator Duane Ebanks duanee@kmimediagroup.com Circulation Barbara Gill barbg@kmimediagroup.com Data Specialists Raymer Villanueva raymerv@kmimediagroup.com Summer Walker summerw@kmimediagroup.com
KMI MedIa Group LeadershIp MaGazInes and WebsItes Border & CBRNE Defense
Ground Combat Technology
Geospatial Intelligence Forum
Military Advanced Education
Military Information Technology
Military Logistics Forum
www.GCT-kmi.com
www.GIF-kmi.com
www.MAE-kmi.com
www.MIT-kmi.com
www.MLF-kmi.com
Navy Air/Sea PEO Forum
Special Operations Technology
Tactical ISR Technology
U.S. Coast Guard Forum
www.TISR-kmi.com
www.CGF-kmi.com
Border Threat Prevention and CBRNE Response
SPECIAL SECTION:
Integrated Fixed Towers
Subscription Information
Geospatial Intelligence Forum
ISSN 2150-9468 is published eight times a year by KMI Media Group. All Rights Reserved. Reproduction without permission is strictly forbidden. © Copyright 2013. Geospatial Intelligence Forum is free to qualified members of the U.S. military, employees of the U.S. government and non-U.S. foreign service based in the U.S. All others: $65 per year. Foreign: $149 per year.
Border Protector
www.BCD-kmi.com
June 2012 Volume 1, Issue 1
Michael J. Fisher Chief U.S. Border Patrol U.S. Customs and Border Protection
Leadership Insight: Robert S. Bray Assistant Administrator for Law Enforcement Director of the Federal Air Marshal Service Transportation Security Administration
Wide Area Aerial Surveillance O Hazmat Disaster Response Tactical Communications O P-3 Program
www.BCD-kmi.com
Military Medical Military Training & Veterans Technology Affairs Forum
The Communication Medium for Navy PEOs
Corporate Offices
SPECIAL SECTION:
CARRIER ONBOARD DELIVERY OPTIONS
KMI Media Group 15800 Crabbs Branch Way, Suite 300 Rockville, MD 20855-2604 USA Telephone: (301) 670-5700 Fax: (301) 670-5701 Web: www.GIF-kmi.com
Carrier Craftsman Rear Adm. Thomas J. Moore
www.npeo-kmi.com
U.S. Navy Program Executive Officer Aircraft Carriers
Presidential Helicopter O Shipboard Self-Defense O Riverine Patrol Craft Precision Guided Munitions O Educational Development Partnership
www.M2VA-kmi.com
www.MT2-kmi.com
www.NPEO-kmi.com www.SOTECH-kmi.com
Wherever decisions demand intelligence, terrago delivers When mission success depends upon finding and sharing the most accurate, timely and relevant location intelligence, from any source, TerraGo delivers. TerraGo allows organizations to discover, analyze and share powerful geospatial intelligence to the edge ... and back. Find the location intelligence solutions that matter to your mission at www.terragotech.com.
NEWS FEEDS
Social Media
Documents
maps & imagery
Experience the all-new TerraGo at www.terragotech.com
PROGRAM NOTES
Compiled by KMI Media Group staff
Remote Sensing Contract Supports Homeland Security The Geospatial Management Office within the Department of Homeland Security (DHS) has selected four companies to provide geospatial imagery and analysis for real-time intelligence products. The prime contractors on the Remote Sensing Services to Support Incident Management and Homeland Security contract—Aero-Metric, BAE Systems, Dewberry Consultants, and GMR Aerial Surveys, dba Photo Science—will be eligible to compete for task orders totaling up to $50 million over five years. Under the contract, intelligence experts will acquire, process and disseminate geospatial data and airborne imagery to produce highresolution maps that reflect current environmental conditions. The data will be used to produce real-time intelligence products to support a
variety of DHS missions, to include emergency management of natural and manmade disasters, and possibly security planning for special events. The geospatial intelligence products may also be used to assist public safety and law enforcement with tactical planning and incident response. “Our imagery and analysis will provide a new level of situational awareness for first responders, helping them to manage emergencies and safely complete their missions,” said DeEtte Gray, president of BAE Systems’ Intelligence & Security sector. BAE Systems’ Geospatial Operation for a Secure Homeland-Awareness, Workflow, Knowledge program develops hybrid teams of data providers, systems integrators and IT professionals capable of rapidly transforming geospatial data into actionable intelligence.
PEOPLE Rear Admiral Ted N. Branch has been nominated for appointment to the rank of vice admiral and for assignment as deputy chief of naval operations for information dominance, N2/ N6, Office of the Chief of Naval Operations/Director of Naval Intelligence.
Compiled by KMI Media Group staff
within the Office of the Director of National Intelligence (ODNI), has been appointed deputy director of the CIA for science and technology. Kevin P. Meiners, who has been serving in the Office of the Under Secretary of Defense for Intelligence, will replace Meyerriecks at ODNI.
Rear Admiral (lower half) Willie L. Metts will be assigned as director for intelligence, J2, U.S. Pacific Command. He is currently serving as deputy chief, tailored access operations, S32, National Security Agency.
Thomas Kirchmaier
The SI Organization has named Mac Curtis president and chief executive officer. He succeeds Bill Graham, who will serve as vice chairman going forward.
Dawn Meyerriecks, who has been serving as deputy director of national intelligence director for acquisitions and technology
4 | GIF 11.5
within the National Security and Defense Division of Parsons Government Services, a primary business unit of Parsons Corp.
Randy Phillips
Mac Curtis
Dawn Meyerriecks
military affairs. Thomas has been serving as commander, Special Operations Joint Task Force-Afghanistan/NATO Special Operations Component Command-Afghanistan.
Army Major General Raymond A. Thomas III has been nominated for appointment to the rank of lieutenant general and for assignment as associate director of the CIA for
General Dynamics has named Thomas Kirchmaier vice president and president of the company’s Advanced Information Systems business, succeeding Lewis Von Thaer. Kirchmaier was most recently senior vice president and general manager of General Dynamics Information Technology’s Intelligence Solutions division, where he has been since 2007. Joseph Nicholas has been hired as intelligence sector manager
TASC has named Randy Phillips, a former CSC executive, as senior vice president of corporate development and chief strategy officer. In addition, Joseph Pacileo has become vice president of the company’s Mission Solutions business unit, where he will oversee operations supporting intelligence and cyber customers at Fort Meade, Md. Pacileo previously served as vice president of cyberoperations for ManTech International.
www.GIF-kmi.com
Streamline your workflow from beginning to end with unparalleled search functionality, exploitation capabilities, and product creation for commercial and defense industries. Discover your data with GXP Xplorer. Search multiple data stores across an enterprise with a single query to locate imagery, terrain, text documents, and video. Exploit data using SOCET GXP® to create geospatial intelligence products with advanced feature extraction tools, annotations, and 3-D visualization for planning, analysis, and publication. Our Geospatial eXploitation Products give you the power to deliver actionable intelligence, when it counts.
Imagery courtesy of DigitalGlobe
www.baesystems.com/gxp
GXP XPLORER AND SOCET GXP. MAXIMIZE YOUR PRODUCTIVITY — FROM DISCOVERY TO EXPLOITATION.
PROGRAM NOTES
Compiled by KMI Media Group staff
Seeking Location Data Without the Satellites Deployment is getting underway this summer at White Sands Missile Range (WSMR), N.M., of a new type of positioning system that, developers say, could reshape precise location technology and reduce reliance on the decades-old GPS system. An Australia-based company called Locata recently announced that the Air Force 746th Test Squadron (746 TS) had awarded TMC Design Corp. a contract to install and integrate the Locata Non-GPS Based Positioning System (NGBPS), providing the unit with highly accurate “truth data” covering a large proportion of WSMR. The Locata system will give the 746 TS an entirely new and enhanced capability, allowing them to evaluate the operations of mobile and airborne position, navigation and timing equipment in GPS-denied environments. The operational installation of the NGBPS system will consist of both fixed and mobile command and control nodes, and a network of LocataLite transceivers deployed on WSMR. The Locata NGBPS system is capable of operating in combination with GPS, or completely independent of GPS, depending on mission requirements. As the GPS Test Center of Expertise and lead test organization for evaluating GPS user equipment and GPS-based guidance and navigation systems, the 746 TS recognizes the NGBPS as a critical component for the realization of the squadron’s new UltraHigh Accuracy Reference System for the increasingly demanding test and evaluation of future navigation systems for the Department of Defense. The key to the Locata system, explained Chief Executive Officer Nunzio Gambale, is that it avoids the elaborate infrastructure, including atomic clocks, orbiting satellites and a complex ground
structure, needed to operate GPS. Instead, an array of video-cassette-sized devices strategically placed around an area can achieve the exquisite time synchronization needed to provide location data, thus creating “a system that breaks the paradigm of having to have satellites in space to give you this capability,” Gambale said. Unlike GPS, which provides a weak signal relatively easily subject to interference, the Locata system can provide a signal so strong it is virtually impossible to jam, Gambale said. “One of the weaknesses of GPS is that it is a 1970s technology now trying to cope with 21st-century electronics,” he noted. “The problem with the enormous dependency on GPS is that if anything goes wrong, what is the backup? If you want to provide reliability and redundancy for modern applications, you can’t guarantee it with satellites.”
NGA Eyes Transformation of Aeronautical Data Model As part of its mission to provide time-sensitive aeronautical and navigation data to support Department of Defense operations, the National Geospatial-Intelligence Agency has asked industry for advice on ways to transform geospatial information from a product-centric model to an integrated data-centric model. The goal of the effort, still in its early stages, is to develop and integrate the DoD Geospatial Terminals Operations so that NGA can provide 6 | GIF 11.5
worldwide aeronautical safety of flight information and data to support operational needs of DoD flight crews. The goal is to automatically render DoD Terminal Instrument Approach Procedures (IAPs), Standard Instrument Approach Procedures (SIDs) and Standard Terminal Arrival Charts (STARs) from NGA source holdings. The capabilities desired by NGA include generation of new IAPs, SIDs and STARs, maintenance of the
plates, retrieval and storage of data geospatially, generating the renderings textually and in Open Geospatial Consortium formats, all within the NGA prescribed 28 day production cycles. In seeking industry input during a nowclosed comment period, officials emphasized ideas that could result in limiting unnecessarily constraining requirements, realizing cost savings, or highlighting potential technical or contractual problem areas. www.GIF-kmi.com
Commercial providers develop new products and find innovative ways to help customers manage torrents of data.
By Harrison Donnelly GIF Editor
We have a really positive outlook and see incredible potential,” she As it copes with tightening limits on its military and intelliadded. gence markets, the commercial remote sensing industry is rein“Government budget cuts are clearly having an impact across venting itself by developing new products and finding new ways to the whole industry. Regarding data sales, I think these cuts are hithelp customers manage the torrents of data being brought in by a ting providers of very high resolution harder, as this is the segment variety of advanced sensors. in which the government has traditionally spent more money in Leading companies in the field, which include U.S.-based the past. But nevertheless, the U.S. government will keep being the DigitalGlobe as well as international competitors such as Astrium number one customer for data in the country. On the other hand, Services, GEO-Information and RapidEye, acknowledge the chalthis is in a way changing the rules of the game now, and pushing lenges created by flat or declining government budgets for satellite those players that used to make most of their sales revenue out of imagery. But they are working hard to develop new markets, adding this market to explore new markets with new business models,” value to their imagery products by incorporating analytical services said Clint Graumann, director North America, RapidEye. and, in one case, calling for changes in government restrictions on “Another interesting change we are experiencing in the market the products they can sell. is how the services side of the business is growing,” he continued. The shifts are most obvious at DigitalGlobe, which is complet“Today, the enormous availability of data from multiple sources, ing its combination with former competitor GeoEye. That left it the both archive imagery and new collects, is helping boost the growth only major U.S. company still standing and consolidated its busiof value-added and derivative products. And I think ness with the National Geospatial-Intelligence Agency this is only the start. Growing archives, more and under the EnhancedView contract, which provides a more missions imaging the earth every day, and a solid base of about 35 percent of company revenue. more efficient access and processing of these data DigitalGlobe’s most prominent initiative is are definitely making feasible applications that a WorldView-3, the first multi-payload, super-spectral, few years ago were not possible.” high-resolution commercial satellite, which is schedGreg Buckman, chief executive officer of uled for launch next year. At Astrium Services, the Astrium Services, offered a similar assessment: emphasis is on the benefits provided by its diverse “What we’re finding with the huge amount of data constellation, which includes the recently launched being collected is that a lot of customers, particuPléiades satellites and the latest versions of its SPOT larly the U.S. government, are being flooded with series of orbiters. RapidEye, meanwhile, has rolled out Marcy Steinke pixels. While this isn’t new, we’re hearing from its Mosaics imagery products, which are seamlessly our customers that they like having all the data that our satellites created to cover a country or a region with minimal cloud cover and bring. But what can we bring them that is new and different to help provide a high-quality, up-to-date base map product. us achieve our missions on the ground to support the warfighter? “I’m very excited about where commercial imagery is going in They aren’t looking for data so much anymore as for solution sets the market,” said Marcy Steinke, a retired Air Force colonel who and services.” currently serves as senior vice president of government relations at DigitalGlobe. “I know that when I talk to both former colleagues and others at the combatant commands and elsewhere, if they Change Detection know they are using commercial, they love it because it’s shareable, and that creates such benefits in working with allies. At RapidEye, one important focus is on change detection, lever“Or they might not know it’s shareable, which is one of our aging the company’s worldwide archive of imagery. “Change detecchallenges for the future. We don’t really care if they know it’s comtion programs can really take advantage of this capacity, and this mercial, except to help make sure they maximize the benefits that is one of the main lines we are currently developing in the intellithey already have available through the EnhancedView contract. gence market, along with some partners. Land cover mapping with www.GIF-kmi.com
GIF 11.5 | 7
regular updates, potentially more than one a year, over broad areas is another product/service line we are currently exploring in this field,” said Graumann. One of the key advantages of RapidEye Mosaics, he added, is how recent and homogeneous in time the source imagery used is. The latest version of the North Korea mosaic, for example, was produced with data collected over a period of just three months. “The other main differentiator is how reliably we can collect data over difficult-to-image areas due to cloud persistence, such as the tropical belt,” Graumann said. “Our collection times are very good compared to other data sources. Right now, at this resolution level, our data is the only option to produce such current and homogeneous mosaics over whole countries or large areas. It is the perfect solution for those looking for ready-to-use datasets over large areas that put together recent data collected over short periods of time.” At Astrium Services, executives are heeding demands from their military and intelligence customers for mission support in the form of the capability to cover large, countrywide areas on a regular basis, and to track both feature and artificial change detection on the ground. “If I’m looking at Syria, for example, and I can get an updated map every 60 days at a very low cost and with rapid turnaround, and can find new roads and buildings without having to use a lot of analyst time, that would be a game changer in the marketplace. That is the strategy we’re following at Astrium Services,” said Buckman. “There are two components that drive that strategy, revolving around SPOT 6 and SPOT 7,” he said. “SPOT 6 has a unique resolution in the marketplace—1.5 meters. The old SPOT system wasn’t very strong in feature and change detection on the ground, but with the new SPOT 6, we’re finding a lot of sub-meter users are moving to a meter and a half because of the greater collection and intelligencegathering capability of the new satellite, which has only been commercial for a few months. “What we’re able to do is to collect complete country coverage extremely rapidly, and do low-cost feature detection and artificial human change detection on the ground with extremely highly accurate results.” The Pléiades 1A and Pléiades 1B satellites, which operate as a constellation in the same orbit, phased 180 degrees apart, also offer unique capabilities, according to Drew Hopwood, Astrium Services technical sales engineer. “We’re the only player in the market that has two identical very high resolution satellites. What it means for us is that we can collect a lot more data more quickly. Our clients can benefit from rapid collection of large areas, as well as daily revisits,” he said. After the recent tornado in Oklahoma, for example, the company was to acquire an image over the affected area for the four days that followed the event. Along with expanding its international and commercial markets, DigitalGlobe is looking to strengthen its government business by offering new services and encouraging demand for existing ones. “For overall growth in the future, we’re also looking to move into the information space, and be a partner with the government in that realm as well,” said Steinke. “We have a variety of initiatives right now. Because of our amazing collection capability and unrivaled archives, we are looking at being the geospatial big-data provider worldwide. “One of the major benefits from the combination has been the analytic side that came in from GeoEye. The former DigitalGlobe 8 | GIF 11.5
had a great group of analytics people with lots of talent, but GeoEye had done a lot with analytics. So we’re really excited about the combination. We’re already working on human geography and crowd sourcing issues,” she said, noting the recent acquisition of Tomnod, a company that analyzes imagery through crowdsourcing software, and the hiring of its co-founder, Shay Har-Noy, as director of research and development. The company is also trying to expand awareness among potential government users of the benefits available under the EnhancedView contract. “We’re not content to just have the contract, but want to make sure it is being used to the maximum extent possible,” Steinke said. “Our contract with NGA provides to almost all of DoD and other parts of the government, so we’re looking at education and training, going out to combatant commands and NGA, to make sure that people are aware of the capabilities that are out there. We’re finding that a lot of them were not.”
Image Restrictions Looking ahead, a key item on DigitalGlobe’s agenda is addressing U.S. government restrictions on the imagery it can sell to other customers, which do not apply to other companies in the world market. “A critical issue is reducing resolution restrictions in the future. Now, we’re restricted from selling anything less than 0.5 meter resolution commercially. When that was put in place nearly 20 years ago, that probably made sense from a national security perspective. But now, as in other areas, technology has far surpassed legislation, so you have a worldwide aerial market selling imagery at resolutions down to 5 centimeters. All we’re asking is to be put on an equal footing,” said Steinke. “We feel that we’re the world leader now on quality and capability, and we want to keep that position. Many companies and countries are close on our heels, and there is nothing blocking them from moving beyond us,” she added. RapidEye’s priorities include improving change detection approaches by providing the necessary high-resolution, multitemporal imagery over short periods of time. “Now we are ready to go with subscription programs to meet the needs of such change detection programs by granting access to our extensive and recent archive, as well as new data collected on-demand with our five satellites. In addition, we are currently working on cloud-based solutions to streamline discovery, access and processing of large amounts of data, making all these three tasks much more efficient,” said Graumann. Buckman summed up his company’s approach this way: “We see tremendous growth opportunity in this market. We realize that there are budget cuts, but that paves the way for new and innovative products that are able to help service U.S. government customers. We are the only remote sensing company that has a true satellite constellation in the sky—twin satellites and those that are in complementary orbits that can image in multiple resolutions and technologies, including radar or optical, anywhere on the globe, every day of the year.” O For more information, contact GIF Editor Harrison Donnelly at harrisond@kmimediagroup.com or search our online archives for related stories at www.gif-kmi.com.
www.GIF-kmi.com
Global Enhanced GEOINT Delivery Based upon the NGA
ENhANcEdViEw award, commercial
imagery access is easy. The Global Enhanced GEOINT delivery (G-EGD) program provides NGA and its customers with current unclassified high-resolution imagery in support of operational planning, emergency response and situational awareness.
» READY TO USE CURRENT IMAGERY
» “RAPID RESPONSE” IMAGES OF H I G H - P R I O R I T Y L O C AT I O N S
» ON-DEMAND ACCESS
Have a .mil or .gov e-mail? You could be eligible to access current global imagery. US / MEXICO BORDER
Go to:
digitalglobe.com / egd
XT5-EGJL2
Video Enhancement 10 | GIF 11.5
www.GIF-kmi.com
By improving image quality, new technologies are making full motion video more valuable to analysts, commanders and warfighters.
By Peter Buxbaum GIF Correspondent
Full motion video (FMV) technology is becoming an increasingly important tool in military and intelligence operations, and efforts to develop technologies that make that video more useful and valuable to analysts, commanders and warfighters are keeping pace. FMV often suffers from quality issues that diminish opportunities to collect actionable intelligence. Problems like low resolution and lack of platform stability, as well as environmental issues such as the presence of dust or haze, compromise the value of the video output. Industry has been working on solutions to deal with the problem of enhancing video in real time, typically involving both software and hardware. The real-time piece requires enormous computing capacity and complex algorithms. In order to deliver instantaneous results, companies have found it advantageous to make use of parallel computing capabilities that break down tasks into subparts, each of which is processed separately before being reassembled. Parallel computing requires specially written software that can process the video quickly enough to produce the desired results, as well as hardware that can accommodate that type of processing. “Organizations responsible for working with video are finding no shortage of raw content,” said Robb Mott, vice president, military and intelligence solutions at Intergraph. “There are more sensors and more systems, and they are all becoming more reliable and more economical to deploy. The problem is how to create an effective environment www.GIF-kmi.com
for the operator or user to get the most value out of the video. There are most often some type of defect or artifacts with video. It may be jittery or fuzzy, or something may be obscuring a clear view.” “Camera motion is one challenge to producing high-quality imagery on a mobile platform,” said Jason Wade, vice president of product marketing and sales at Z Microsystems. “The quality of the video imagery can also be compromised by poor environmental conditions, data link degradations and bandwidth limitations. Atmospheric factors, such as poor lighting at dawn, dusk or nighttime, and adverse weather, including sandstorms and variable clouds, can obscure important details.” “There is also the issues of the stand-off distance between the sensor and the subject, compression artifacts due to low-bitrate links, and signal degradation due to multiple transcodings or analog-digital conversion in the signal path,” added Sean Varah, chief executive officer of MotionDSP.
Quick Answers One of the most important challenges with FMV is to be able to process the video without latency. “Answers need to be generated much more quickly and decisions need to be made much more confidently within a shorter time frame,” said Mott. “What is critical for a solution is to include some type of algorithm that performs enhancement of the video in real time.”
Rob Mott
GIF 11.5 | 11
“They can see de-hazed and stabilized Another key point is that the origivideo playing in real time side by side nal video in some cases needs to remain with the original video. The system prointact for legal or intelligence reasons. duces, not just video enhancement, but “Enhancing video on the fly means allowit can work in a multi-int environment to ing users to view the video without preincorporate geo-referencing.” processing that can delay their activities,” said Mott. Some signal problems affecting video Hardware Advances can be fixed by the users themselves, noted Varah. “They can plan a cleaner Advances in hardware help MotionDSP signal path between the sensor and the and Z Microsystems process video in real analyst. We see several customers doing time. this now. But greater standoff distance “Continued advancement in graphic between the sensor and the target means processing unit (GPU) architectures, such there is going to be more haze in the as NVIDIA’s Kepler architecture, has proimage.” vided leaps in computational capacity The platform and environmental while using less electrical power,” said issues require the application of realVarah. “This has allowed us to double or time image processing algorithms. triple the number of video streams we can “Algorithms analyze how the neighborprocess on a single GPU card. We can pack ing frames in a clip are related to one more processing in a smaller footprint. another,” said Varah. “It is “Two years ago we were possible to recover the best lucky if we could process a available image informasingle video steam on one tion from multiple frames video card,” he added. “With and reconstruct each the new architectures we can frame of video with this process six streams of staninformation to enhance dard definition video and the appearance and qualthree streams of high-definiity of the video.” tion video at once on a single “The only way to ensure card, so the capabilities have a good image quality is to gone way up.” Sean Varah give viewers the ability to Using conventional comadjust the picture for their needs,” said puter processing would result in between Wade. “The best way to accomplish this four and six hours of processing for each is to bring real-time video enhancement hour of video, according to Varah. “This to the tactical edge. Field programmable process takes too much time to enable gate array (FPGA) semiconductors offer information to be viewed in real time, the performance, design flexibility and resulting in potentially outdated intelliresilience needed to build this capability gence data and inaccurate guides for milidirectly into the display controller.” tary action,” he said. “In contrast, the new Intergraph has been improving its generation of GPUs enables MotionDSP’s product, Video Analyst, to enhance Ikena ISR software to process live video in unclear or defective video to create results real time with less than 200 milliseconds that can support decision-making proof latency.” cesses. “The solution is a commercially Exploiting the computing capabilities available product that incorporates a of GPUs also reduces the IT footprint series of video enhancement algorithms,” required for processing video. “You can said Mott. “There are a number of patpack the processing into servers that take ents associated with the technology. The up less space and use less power than the system is also built on open standards, huge workstations that analysts had to which is important for the flexibility it work at before,” said Varah. affords when moving from one mission to MotionDSP recently came up with a another or even from a military to a civilnew de-hazing algorithm that works faster ian context.” with GPUs. De-hazing and contrast issues Users connect to Video Analyst by way brought about by sensor stand-off disof a dashboard with a control panel. “They tance will emerge as a more important work with half a dozen to a dozen factors issue in future military engagements, in at a time to adjust the video,” said Mott. Varah’s estimation. “Our own research and 12 | GIF 11.5
development are addressing new issues that are emerging in anti-access aerial denial environments (A2AD) such as our new de-hazing algorithm and contrast filter for shadow conditions,” he said. A2AD environments contrast with those the U.S. military found in Iraq and Afghanistan, where there was no significant opposition to U.S. forces in the skies. “The U.S. was able to fly anywhere and not get shot down,” said Varah. “In new areas, they may not have that luxury. In denied access environments, we can’t always fly aircraft directly over areas we want to look at.” The algorithms used to de-haze images taken from stand-off positions are both complex and intelligent. “They understand the 3-D perspective, and apply more dehazing to objects further in the distance and less de-hazing to closer objects. This requires the kind of compute power provided by GPUs,” Varah explained. MotionDSP’s contrast filter for shadow is also new. “This involves a different challenge,” he said. “Some images may include the very bright sunlight of the dessert combined with objects such as buildings that are in the shadow. The human eye is more sensitive to light than a video camera is. If you have a vehicle or an individual of interest in shadow, you won’t see it in the image. We developed a locally adaptive algorithm that applies itself to dark areas of shadow and not to other areas of the image. As a result, you can see much clearer results now than you would only two years ago.”
Analytic Algorithms Algorithms developed for video analytics and ported to GPUs now can use a tool called a visual moving target indicator (VMTI), to allow for the accurate detection and tracking of objects on the ground from manned or unmanned airborne platforms. “We were able to develop these capabilities thanks to the greater compute capacity now available to us on the new GPU cards,” said Varah. “Earlier algorithms produced a lot of false positives. They erroneously reported movement when there was none. False positives are very distracting to the analyst, so reducing them is very important.” Motion DSP has likewise been able to develop a feature detector that is much more accurate than those previously used. www.GIF-kmi.com
The FPGA technology used by Z Microsystems has been around since the 1980s. It has been utilized in the health care market and other commercial applications, but was only recently applied to the real-time processing of FMV. “What is different about what we are doing now is that we are providing a real-time capability in ruggedized hardware that can be used in fielded environments,” Wade said. FPGAs allow the use of sophisticated algorithms that can enhance video in real time. “They are computational workhorses that are well suited to military video display controller applications,” said Wade. “FPGAs can withstand harsh environments and meet military requirements for ruggedness, temperature tolerances, reliability and product lifespan.” FPGAs are also reprogrammable, so that a display controller can be adapted to changing video standards, or special mission requirements. “Once deployed, FPGA-based display controllers can be field upgraded to add additional features and new image enhancement algorithms. It is possible to swap one algorithm for another as the need arises on our open source hardware,” he added. With the ability to compute in parallel rather than serially, FPGA hardware facilitates parallel processing in Z Micro’s products. “As a video stream comes in, there is an intensive amount of pixels that need to be analyzed,” said Wade. “It can’t be done pixel by pixel. Conventional processor technology does not offer the performance necessary to keep up with the demands of FMV at up to 60 frames per second.” When image enhancement algorithms are written using parallel processing techniques on an FPGA, it is possible to dramatically enhance ISR video in real time, Wade noted. “Of the many types of image enhancement algorithms, spatial convolution kernel filtering produces the most dramatic results. A convolution kernel generates a new pixel value based on the relationship between
www.GIF-kmi.com
the value of the pixel of interest, and the values of those that surround it.” The algorithms deployed by Z Microsystems use a large kernel convolution, which means they work on relatively large arrays of pixels at once. One such algorithm, designed to remove haze and enhance image detail, works by mathematically enhancing the imperfect array based on a model of a perfect image. “The technology works backwards,” said Wade, “by stripping noise and image blur while adjusting the intensity of each pixel until the simplest image that fits the real-time data emerges. Because we know how environmental factors distort the image, the distortion can be undone. The algorithm is continually applied to the image until it is as close to the perfect image as possible. It can also take an image that has both bright sunlight and a dark cave in same image and enhance both what is inside the cave and what is in direct sunlight.” Z Micro’s solution was beta tested a year ago and is out in the market as part of its Real Time Enhanced Video solution, where it has attracted several military customers. Earlier this year, the company was selected by the Naval Surface Warfare Center to supply its Intelligent Display Series for use in the Small Tactical Unmanned Aerial System program onboard amphibious ships. “It is best suited to environments that make it difficult for the human eye to discern objects,” said Wade. “The algorithm takes the human subjectivity out of it and pulls in details the human eye might not be able to see. The feature can be turned on and off with the push of a button. The Navy is using it on UAVs in fog, haze and dust. Other military customers are looking to use the system for underwater video enhancement.” The company is also in the process of rolling out its new ZBox product, which combines algorithm requirements with storage and processing requirements. “This is one box that is an all around solution for military environments in both ground stations and command and control nodes,” said Wade. Z Microsystems continues to identify ongoing algorithm needs, which it then develops or acquires from others. “We can then stitch those algorithms together and incorporate them into our system to get better results,” he added.
Activity-Based Intelligence The software and hardware developments that make the new generation of video enhancement possible also are facilitating the growing interest in activitybased intelligence, noted Mott. “Analysts are using video in a geospatial context to look at changes over time. This allows analysts to discern patterns of activity to solve problems that activities-based intelligence presents.” Intergraph’s Video Analyst lets users overlay different sources of intelligence on the video to produce more meaningful intelligence. For example, a user may be tracking a vehicle on a road. Overlaying a map can provide an analyst with more valuable situational information. “It is a powerful way to take isolated or compartmentalized geospatial data and bring them together in a dynamic environment without a lot of pre-processing,” said Mott. “With live links to video and GPS, users can track vehicles in a really rich environment for making better decisions.” MotionDSP’s research and development is focused on three areas. “We are seeking advances to our core video improvement algorithm suite to achieve better quality, greater automation and richer feature set,” Varah said. “We are striving for true ISR video analytics, in real time. Our new VMTI capabilities are tuned specifically to ISR sensors for detection of extremely small objects, with high false-positive rejection, all in real time, on both high-definition video, and wide area motion imagery data sets. We are working on integration with all of the major FMV ground control station video systems as well as those embedded at the sensor directly on board the ISR platform.” As the recent U.S. military missions in Southwest Asia wind down, video enhancement technologies can be put to use for other missions, noted Mott. “They can be used for humanitarian assistance and other relief efforts. They can be used to survey coastal areas after a hurricane or others areas damaged by earthquakes. These technologies can bring great value to folks who support those types of missions.” O For more information, contact GIF Editor Harrison Donnelly at harrisond@kmimediagroup.com or search our online archives for related stories at www.gif-kmi.com.
GIF 11.5 | 13
INDUSTRY RASTER Defense Intelligence Agency Orders Enterprise IT Support Northrop Grumman has been awarded an Enterprise Application Development and Integration Support task order with an estimated value of $318 million over four years. The task order is provided through the Solutions for Information Technology Enterprise contract from the Defense Intelligence Agency (DIA). This task order will support the full software development life cycle of DIA’s consolidated and efficient enterprise-centric information technology architecture. It also includes the design, development, integration, testing and deployment requirements, tasks and technical support services to DIA’s Chief Information Office. Stephanie Trumpower; stephanie.trumpower@ngc.com
GIS Cloud Service Aids Emergency Response Esri and Witt O’Brien’s LLC have built a new GIS cloud service that will increase an organization’s capabilities to prepare for, respond to and recover from incidents. For the first time, emergency management personnel can visualize an integrated incident command system (ICS) through a common operational picture (COP) and systematically command, control and coordinate multiagency emergency response. First responders, government agencies and private/commercial industries will have the combined ability to respond faster, control resources better and generate timely incident reports accurately. The new COP will improve their vital communication and coordination capabilities and give them complete enterprise visibility. The combined technologies in the COP simplify the complex process of ICS and increase the agility and capabilities of managers dealing with disasters of various sizes and scopes. Witt O’Brien’s CommandPro automates emergency management processes. Esri’s ArcGIS Online platform provides users with access to fit-for-purpose maps and workflow-enabled apps. The integrated solution aligns real-time data with the systematic approach of ICS. Lynn Holmes; lholmes@wittobriens.com
API Brings Mapping Power to Application Development The Google Maps Engine API allows developers to bring the power of Maps Engine into their own applications. Maps Engine lets organizations use Google’s fast and reliable cloud infrastructure to layer their data on top of a Google Map and share their custom-made Google Maps with employees, customers or the publicat-large. At the push of a button an organization can share their maps internally or publish them on the web, making it easy to access their data from anywhere. The API provides direct access to Maps Engine for reading and editing 14 | GIF 11.5
spatial data hosted in the cloud. With the API, organizations can develop on any platform, including web, Android, iOS and server-to-server, and build applications like store locators, crowd-sourced maps or crisis-response maps. Google Maps Engine is changing the way enterprises create and use maps—replacing complex GIS technologies with accurate and comprehensive Google Maps-based tools.
Data Management Solution Supports National GEOINT System The National Geospatial-Intelligence Agency has awarded BAE Systems the iSToRE XP contract to provide advanced data management capabilities in support of National System for GeospatialIntelligence (NSG) sites and users. The iSToRE XP software solution is built on BAE Systems’ commercial product, GXP Xplorer, which enables analysts to easily access their local data and connect to remote geospatial data stores and libraries. GXP Xplorer searches through unlimited amounts of data to identify and catalog images, terrain, features, videos and documents on local networks or across an enterprise. The software operates on a variety of hardware platforms scaling from handheld devices and ruggedized laptops to enterprise servers and virtualized cloud environments. Under the contract, BAE Systems will provide GXP Xplorer licenses for deployment during NGA’s transition from legacy image product libraries to commercial-based information storage capabilities. The contract includes options to operationally field several additional licenses to the NSG through 2016. The licenses would support worldwide operations, including joint task force headquarters, intelligence centers and forward-deployed warfighters. Charles Ratzer; charles.ratzer@baesystems.com
Tools Improve Motion Imagery Processing 2d3 Inc. (dba 2d3 Sensing) has announced the launch of Reticle FMV metadata improvement and georegistration in the latest release of its motion imagery processing, exploitation and dissemination product suite. The suite includes TacitView 3.2, Catalina 3.1 and Tungsten 3.2. Integrated within Catalina, Reticle improves geospatial data for use in mapping and targeting, correcting metadata errors typical of aerial full motion video platforms. Reticle comes in three levels of building-block capabilities: Metadata Optimization (MO), Real-Time Georegistration (RT) and Offline Target Coordinate Improvement (TGT). Reticle MO is designed to clean up and normalize noisy metadata in an FMV feed or file, Reticle RT furthers improves on Reticle MO by using reference digital elevation models and imagery to minimize geospatial bias, and Reticle TGT takes a user-selected point from full motion imagery frames and computes the geoposition of that point along with estimates of uncertainty. Each step in Reticle improves upon, and makes use of the techniques in the previous step. With these releases, 2d3 Sensing is making the next generation of motion imagery processing available today—providing more tools than ever to go from imagery and information to intelligence. Danny Proko; danny.proko@2d3.com
www.GIF-kmi.com
Compiled by KMI Media Group staff
Target Detection, Video Exploitation Systems Linked General Dynamics Mediaware and Sentient Vision Systems Pty Ltd. have created a technical partnership that combines Sentient’s Kestrel Land and Maritime target-detection software with General Dynamics Mediaware’s D-VEX next-generation tactical video-exploitation system. This plug-in feature provides surveillance system operators and analysts with improved real-time situational awareness and strengthens post-mission forensic analysis and intelligence reporting. D-VEX is a video-exploitation system that captures and manages full-motion
Software Offers Improved Handling of Large Data Sets Overwatch, an operating unit of Textron Systems, a Textron Inc. company, has released version 3.2.3 of its RemoteView geospatial analysis software, providing imagery and geospatial intelligence analysts improved handling of large data sets and access to Overwatch’s valuable GeoCatalog Workgroup extension. This geospatial analysis tool for intelligence exploitation now offers native 64-bit processing support for Windows 7 64-bit and Windows XP 64-bit computers, delivering fast and efficient handling of the large data files analysts are increasingly required to assess. Overwatch’s GeoCatalog Workgroup software extension enables RemoteView users in small analyst teams to easily share and manage geospatial data and databases on a shared local network. It offers database and file management capabilities to add and remove cataloged data as well as associated files including images, shape files and video. GeoCatalog Workgroup also provides automated ingesting and monitors directories for incoming data. Kevin Ortiz; kopitz@overwatch.textron.com
www.GIF-kmi.com
video, providing operators with intuitive tools for enhancing, streamlining and analyzing live and recorded video. When coupled with Sentient’s Kestrel real-time automatic detection software, which allows users to identify small, hard-to-find moving targets in electro-optical and infrared aerial live video streams, mission operators and analysts can quickly transform raw video data into actionable intelligence. Leveraging the companies’ extensive experience in advanced full-motion video, the combined systems simplify the post-mission forensic analysis of video. This product collaboration with Sentient’s automated detection and D-VEX’s intelligent video technology enables customers to extract significant
value from the large volumes of full motion video captured from sensors deployed in the field. Jessica Howe; jessica.howe@gd-ais.com
Mobile Devices Linked to Computer-Aided Dispatch System Mobile Responder from Intergraph gives personnel in the field access to Intergraph’s computeraided dispatch system on a mobile device. Accessible on Apple iPads, iPhones, and Google Android smartphones and tablets, Mobile Responder can be used as a standalone unit or as an extension to Intergraph’s Mobile for Public Safety software. Because smartphones and tablets are cheaper to deploy than laptop computers, Mobile Responder enables agencies to provide more field personnel with the technology and reduce costs. Mobile Responder takes advantage of the GPS capabilities in smartphones, and provides police, fire and emergency management personnel access to critical information outside of response vehicles. For example, firefighters can take a smartphone or tablet into a building to gain situational awareness, or the location of police officers can be instantly known while they are on foot at an incident. The easy-to-use software streamlines processes and supports intuitive workflows, ensuring accuracy. With Mobile Responder, organizations save time and money with fast deployment, shorter training times and simple updates.
Tactical Imagery Solution Available on Mobile Platform ITT Exelis will make its Jagwire solution for the management and dissemination of tactical imagery and video available on a mobile platform. Jagwire Mobile allows users, such as soldiers operating in environments where bandwidth is limited, to capture, process and rapidly access imagery. The Jagwire Mobile solution offers much of the same functionality of the Jagwire enterprise version, which is important for users in bandwidth-constricted or hostile areas. Taking advantage of advanced data compression techniques, Jagwire Mobile allows video and image capture as well as processing into the Jagwire repository, where the imagery is indexed and becomes searchable by other users on shared networks. The mobile version of Jagwire supports Android 4.1.1 and higher. Support for iOS 6 and HTML5 will be available later this year. Irene Lockwood; irene.lockwood@exelisinc.com GIF 11.5 | 15
Budget restraints are the new normal. so why keep Buying access to space the same old way?
Putting your payload on our next launch gives you reliable access to space while mitigating the effects of federal budget cuts. Take advantage of our proven track record, regularlyscheduled launches, and significant cost savings. And make your overall space architecture more resilient by disaggregating assets. The Australian Defence Force did just that when they shared an Intelsat launch last year for a UHF payload. They’ll save $150 million over 15 years. How about you?
Talk to one of our hosted payload experts. Visit spacedelivered.com.
Future Technologist
Q& A
Supporting Longer-Range, High-Risk, High-Payoff Research Dr. Peter Highnam Director Intelligence Advanced Research Projects Agency Dr. Peter Highnam was named director of the Intelligence Advanced Research Projects Agency [IARPA] in August 2012, after joining the agency in February 2009 as the office director for incisive analysis. Prior to IARPA, he was a senior adviser in the National Institutes of Health and then in the Biomedical Advanced Research and Development Authority. From 1999 to 2003, Highnam was a program manager with the Defense Advanced Research Projects Agency [DARPA], with programs in electronic warfare and airborne communications. Before joining DARPA, he worked for more than a decade in applied research in industry. He is a co-inventor on three patents in commercial seismic exploration and holds a doctorate in computer science from Carnegie Mellon University. Highnam was interviewed by GIF Editor Harrison Donnelly. Q: After more than a year as director and acting director of IARPA, how would you describe your vision of its mission for the intelligence community? A: IARPA has an interesting role, in that we are chartered to do longer-range, high-risk, high-payoff research, while staying away from operational activities. It has been in business for about five years, and I think that we’ve really hit our stride. We have good relationships with just about every component of the intelligence community. What we do here is to involve the potential end-users, or transition partners, in the definition of a program, and they participate in all aspects of our efforts to find research teams to work on our programs. We try to give them ownership before the program even exists to ensure that what we aim to produce and the results of our programs will be of use to them—not necessarily today, but, if the programs succeed, when the results come out in three to five years, which is the typical length of a research project. So it’s not waiting three to five years and then handing over the research results, but letting them see what is working and not working, and then taking those ideas into their own acquisition or engineering processes. We work very hard on that. Q: In broad terms, what would you say were the most important advanced research needs of the IC? A: The IC is diverse and has many different needs. The way we approach this particular challenge is with broadly structured program offices. The first office is Smart Collection. As you collect more data, how do you collect the right data, the data of highest value? Incisive Analysis is the second. We recognize that there will always be more data than you can deal with—it’s always going to be ‘noisy,’ with an increasing number of types. “Big data” is a buzzword that’s now www.GIF-kmi.com
starting to fade, but it’s been the norm for the IC forever. So one of our research offices is devoted to dealing with the massive amounts of data, and helping analysts to make more sense of it. In a strong sense, in the IC we get information to decision makers in a timely manner, and information is touched all along the way by analysts, who bring experience, knowledge and other information to increase the value of that information, until it gets to the person who has to make the call. There are many opportunities in that sequence for specific types of technology, as well as to look at the general topic of humans making judgments. The third research thrust is about the fact that we can’t assume that the networks on which we work are secure. How do you deal with that? This is the Safe and Secure Operations office. A program in the analysis office provides an example where you can make the distinction between our work in the IC and elsewhere. Some people who work on video analysis frequently worry about video from overhead cameras being flown over certain locations. In those situations, it’s our camera. In IARPA’s Automated Low-Level Analysis and Description of Diverse Intelligence Video [ALADDIN] program, however, we don’t assume that we control the camera. There’s no production quality associated with it, and you don’t know what’s in a video clip until you look. How do you provide technology that can analyze this open source video in such a way that I as an analyst can pose a query? I’m looking for events of a certain type, and here are some video examples. The system will then produce the clips in the larger collection that appear to have instances of those events, and give me some evidence as to why that might be true, and why I should now spend my time as a human analyst looking into those particular clips. GIF 11.5 | 17
The program builds on a lot of technology that the research community has built up over the years. It’s a challenging program that we’re making great progress on, and it’s unclassified. One thing I really like about it is that each year we sponsor a competition via the National Institute of Standards and Technology, which lets anyone compete with our teams. We’re not relying on just us saying this is good science, but we’re doing a large open competition against the best in the world. So we understand once we’ve got something how well it works and doesn’t work. We’re proud of that.
we’re supposed to do, and that means making the best calls you can make—re-prioritizing all the time, never getting hung up on a particular technology or approach, and always being willing to change. That’s the responsibility that comes with that charge. Another part of the answer to your question is that there is one DARPA for defense. The original ARPA was created in 1958, so it’s been around for a while. DARPA has a similar mission for the defense world, and having served there, and watched it from the outside as well, I think that they’re doing very well.
Q: Why have a single advanced research agency for the IC, given that it comprises 16 different agencies with widely varying strategies and priorities?
Q: Your past career includes service with DARPA. Since readers may be more familiar with that agency, how is it similar and different from IARPA, beyond the fact that one serves DoD and the other the IC?
A: There are many science and technology activities across the IC. But IARPA is an organization that is focused on longer-term ideas. We don’t do the near term—the low-hanging fruit. Even recognizing that those problems are incredibly difficult, they have to be solved right now. We have the luxury and the responsibility to look at the longer stretch—three to five years or longer—and focus there. That requires discipline, and people who know what we do, how we work and how we’re chartered understand that and recognize it. We were created as part of the Office of the Director of National Intelligence [ODNI], and I think that’s why we’re here, because the ODNI is a place where an R&D organization you can stay focused on the future. I view this as a huge responsibility, because that’s the charter that we have and what
Adapting Technologies to Enhance the Efficiency of Tactical Operations August 26 - 28, 2013 | Washington, DC Metro Area Don’t miss out on mission critical requirements for tactical ISR Attending will allow you to: ✓
Support information dominance and increase the utility of actionable intelligence for unparalleled threat detection
✓
Explore the latest capabilities and techniques to overcome information sharing challenges
✓
Discover the most recent mission critical requirements for intelligence analysis throughout the IC
✓
Better understand the community’s acquisition priorities and how industry can partner for overall collaborative success
www.ActionableIntelSummit.com 18 | GIF 11.5
A: In both cases, the success of the institution relies on the program managers, who are the critical elements. They have the ideas and they drive the programs technically and programmatically at both places. Having been a PM at DARPA and an office director here, and now privileged to serve as director here, I’ve seen this type of organization in multiple roles. And, it really is about the PMs. Like everything else, DARPA has changed over time, and my view of DARPA is primarily from when I was there, which was from 1999 to 2003. The PMs really are critical. When candidates interview at IARPA, they come with a resume, but also they have to come with a program idea, and be able to pitch that and go through a process that we use to understand if the individual can structure a program and run it. You’re taking taxpayer money, looking ahead three to five years, and saying ‘I think we can get from this capability here to another one there, changing a technical field.’ You have to define where you want to go, how you’re going to measure it, and how you’re going to do the competition, which for us is primarily full and open competition, to find the best research teams to work on it. Once it’s running, you are responsible for finance, programmatics, technical performance, making sure that people in the IC know what you are doing, how well it’s working and not working, and to effect the transition of the science and technologies that result. It’s a hugely exciting role, and having been a PM, I think you get enough freedom to make the calls that you need to make within the programs that you are running. It’s much more like being in industry, which is where I came from before joining DARPA, than any other place I’ve been in the government. The second similarity is how program managers justify their new program ideas, through something we call the ‘Heilmeier questions.’ Back in the 1970s, Dr. George Heilmeier was director of DARPA, and he proposed a list of five questions, which we use to structure our new start proposals. They are common to both places. The first question asks for the ‘elevator pitch’ for your program idea. The second asks how well it is done today, the limitations of current approaches, and what is the technical base for the approach that you want to take. ‘Heilmeier three’ asks for your ‘secret sauce’—why do you think you can go from a certain capability to a new level? What is new this time? Don’t just tell me that it is a hard problem, but what you are bringing to it and how you propose to attack that problem. The fourth question concerns metrics—how do we know when we’ve gotten there? When we talk to potential transition partners, we say, ‘If we can do this thing this well, and here are the numbers, will that be a big advantage to you?’ That’s critical—the ‘who cares?’ question. ‘Heilmeier five’ asks how you would structure the program—are there enough research teams out there who could bid? www.GIF-kmi.com
How much will it cost, how do you propose to test it, and how will you bring those metrics back into the program structure, so that as we go through each six months or year of the program, we can tell whether we are on track. It’s critical to be able to do that. Another similarity is that once the program is running, we do regular reviews, looking for things that appear to be too easy, which is rare, or too hard. If it is too hard, why? We scrub everything down, take a look, and make tough calls on research teams that may not be making it, perhaps because their approach, we’ve discovered, isn’t going to cut it. Sometimes we close programs. Q: What is your process for getting input from intelligence agencies about the kind of long-range research they need? A: We ask them. Other than ODNI, there are 16 other components in the IC, eight of which also have a defense hat. It’s interesting to see how often there are problems in common, so that we can go after a program that has multiple interested partners, some of which may not have a science research component of their own. But if they are interested, they can participate and provide advice and feedback. I think it’s important to be open about program successes and challenges. If you run into a problem in testing, it could be that they have been through the same thing already, and can help out. You also want to show them that something isn’t working. You have to show everything, not just the good things. This is why testing and evaluation is so critical in an IARPA program. If we’re going to transition technology in three to five years, we want to be able to say that if you use it in a certain space and in a certain way, it’s going to work this well, and we’ve tested it. If you use it for something else, ‘off label,’ we don’t know. We’re doing things that are important to the IC, if people use capabilities based on our research, we have to be right. This is why independent, rigorous testing is so critical. Q: You have said you are looking for ideas “before they are considered technology.” Why do you emphasize such early involvement, presumably well before the ideas yield practical results? A: If something is already a proven technology, then there are other groups that should be turning it into products and services. This is what venture capital organizations and companies do. If you know the science already, and it’s ‘simply’ a question of productizing, engineering and supporting operations, that may be a hard problem, but there are many commercial organizations that do that really well. Our job is to be upstream of that. Does the science support what we want to do? Can we force a technical field into a space it wouldn’t have gone otherwise, check it out, and from there have those results turned into technology? One example of this is IARPA’s Aggregative Contingent Estimation [ACE] program. The name is IC-speak for crowd-sourced, conditional prediction. How do you take the judgments of a group of people and combine those judgments to produce an overall better judgment? We’re testing approaches by posing questions about real-world events to ACE research teams. Participants vote or invest, as in a prediction market. But in the ACE approach we know a lot about the individuals. A prediction market assumes that you and I are the same, and just have ‘coins’ to invest with. In ACE, we know a lot about you, the participant—we’ve done psychological tests, we know your history of judgments and demographics. If we can combine your judgment on a particular question with those of many other people, and we also www.GIF-kmi.com
know how well they’ve done before, then the overall combined judgment frequently turns out to be more accurate. This is not the case on every question, but the ACE program is doing really well. It’s believed to be the largest study of human judgment. David Brooks wrote an op-ed piece in the March 27 New York Times on this, and after that we had a large bump in terms of people applying to participate. Q: What are you seeking with your recently announced project on high-performance computing? A: We’ve published a request for information [RFI], and we’ll probably have a workshop this summer. The underlying notion is that high performance computing has been driven for decades by complementary metal-oxide-semiconductor [CMOS] technology, which is used in virtually every chip, whether memory or CPU. It’s all coming from the same type of technology. The whole industry is geared up to do that, and given Moore’s Law, it’s been relentless in its performance increases. So over the past few decades, whenever a potential competing technology has popped up, which might need a decade of work to become useful, it may get some funding, but doesn’t go anywhere. If you say you have to wait five years to get better, CMOS has meanwhile jumped ahead. So we don’t have many viable alternatives. People are looking at where CMOS is going to be in a decade, with exascale systems (which is 10 to the 18th power floating point operations per second). At the low end, they are talking about needing 20 megawatts of power to drive that, so electrical power is a constraint, as well as physical size, and some people estimate even higher demand than that. The question we posed in the RFI is about the alternative technologies for high performance computing that perhaps haven’t been invested in, or are at different stages of maturity. We received many good responses, covering everything from DNA computing to optical and other things that you read about in the trade press. Now it’s a question of assessing the maturity of these ideas. Already in IARPA, we had a proposers’ day for a program called Cryogenic Computing Complexity, which is about superconducting supercomputers. How do we work at 4 degrees Kelvin, and take advantage of superconducting properties, where there are essentially no energy costs to moving information down a wire, which is a huge piece of the cost for CMOS systems? This is technology in which there has been almost no infrastructure investment, for example, in the foundries to build niobiumbased devices. Q: What are some of the other projects underway that you would like to highlight to readers? A: One is Trusted Integrated Circuits [TIC]. There has been much in the press recently about the reliability of chips. What we’re doing in this program is recognizing an important fact, which is that 70 percent of the advanced chips in the world are made overseas. All of the newest processes and the smallest feature sizes are available in those places. They have massive volume, which keeps the costs down. That means that if someone wants to fabricate a design in this country, they typically will send the entire design to one of those foundries, where the parts will be made and then brought back. There have been articles in the media about whether something might be in that device that you don’t want. What we’re doing in the TIC program is to try to take advantage of the leading-edge high volume manufacturing processes, but without sending over all of the design. We’re experimenting with asking GIF 11.5 | 19
While IARPA focuses on long-term research into intelligence issues, work goes on throughout the Department of Defense and intelligence communities to develop the full range of intelligence-related technologies. Above, Army Special Forces Sergeant 1st Class Christopher Linnel demonstrates the Lighthouse Sensitive Site Exploitation application he developed through his studies in the Naval Postgraduate School defense analysis program. The app provides a platform for the collection and rapid analysis of intelligence data, and is currently in pilot field-testing.
them to build the transistors, and then ship the wafer back to one of our trusted foundries here, where we will add the metal layers above that, where all the logic is. We’ve demonstrated that we can do that at a certain size, by making the base layer in Singapore and bringing them back and adding the rest. Now we’ll see how far we can push it. This is an interesting and important program, completely unclassified, with some of the best teams in the world working on it. No one has done this, but if we can do it reliably and without performance impact, or a significant cost difference, there are a lot of advantages to this. We also have investments in quantum computing, particularly in two programs, known as Multi-Qubit Coherent Operations [MQCO] and Coherent Superconducting Qubits [CSQ]. In that space, we don’t think we can get from where we are today to someplace really interesting within the three to five year time span we use. But if we look longer, what is the hardest technical challenge that we need to resolve in order to continue with, say, superconducting quantum bits? The CSQ program is looking at that. This is an example of how we’re looking longer term. It’s clearly of interest out there, because people think you can do interesting things. But how do we know what is the right path to follow? So we’re investing in MQCO and CSQ to look at the physics to answer some of the near term questions, to know what might be viable to push further. On the software side, we have a program called STONESOUP. These days, software can be made anywhere in the world. A vendor can integrate software components built anywhere in the world to create a commercial product. But we see every month or so when the major companies put out their security updates that vulnerabilities have been found in their shipped product, and someone has found a way to exploit it. In STONESOUP, we’re not assuming that there is any particular threat in the software we’ve been given. But how do we analyze it and put a wrapper around it so that we can still use it, and not have any vulnerabilities be exploitable? This is not like a virus scanner, where you’re going after specific threats. With this, 20 | GIF 11.5
we don’t know what the vulnerabilities are, but we assume they exist. How can we protect against a large class of them? The win is that we would get to use these software packages much more quickly in our systems than normally, because it takes a long time to vet software before it’s allowed to be used in systems. In the SPAR program, we’re coming up with technology to protect privacy. If you have a database, and I have a query that I want to run against your database, it could be a sensitive query, and your database could have sensitive information. I want to get a response back, but I don’t want you to see my query, and I don’t want you to know what I got back. You don’t want me to have access to anything in your database except the things I’m allowed to have access to. It’s protecting at both ends, and in transit. You can think of a number of instances where that might be useful—you’re not tipping your hand, or giving away information that you don’t want to. Q: Is there anything else you would like to add? A: IARPA is an R&D organization. My observation from working at different places is that, especially in this type of organization, you want the research managers, myself included, to not be there forever. You want to rotate the research people. We do that here, currently with a three- to five-year tour for the program managers and office directors, as well as myself and my deputy. This means that if I have a pet technology, I’m not in a position to keep pushing it if it doesn’t seem to be working. I’m going to leave, and someone else will come in with new ideas and make changes. The same goes for the programs, which run for three to five years and then stop, or sooner if we’re not seeing progress. In any given year, with a four-year typical tour, that’s 25 percent turnover among the program managers. So we’re always looking for program managers who have both a resume and a new idea. Every year, there are new programs being launched as we reprioritize. If a great idea shows up, we make the call and change how we do business. O www.GIF-kmi.com
Analytic tools help military and intel users harvest the rich crop of data available through social media.
By Karen E. Thuermer GIF Correspondent Ever since the 2010 earthquake in Haiti, social media has been regarded as offering real value not only to humanitarian efforts, but also to the military and intelligence communities. “It began to get powerful when used in areas where the humanitarian, military and intel communities have no resources on the ground,� observed A.J. Clark, president of Thermopylae Science and Technology. Today’s social media, with its massive, worldwide participation, offers an unmatched source of information, most of it voluntarily placed in the public domain, about everything from personal pizza preferences to transformative social and political movements. The trick is to find the technology, including location-based geospatial analysis, that can help make sense of www.GIF-kmi.com
GIF 11.5 | 21
all that data, and do so at the dizzying pace that characterizes digital interaction. “The general public has now become aware of the power of social media, and they are now using it purposefully to disseminate information, especially when other options are not available to them,” reported Anthony Stefanidis, a professor with the Department of Geography and Geoinformation Science at George Mason University, and director of the Center for Geospatial Intelligence. Last year, Stefanidis noted, Twitter users posted 400 million tweets daily, doubling the rate of 2011. “These tweets come from an estimated 200 million active users, which is equivalent to the population of Brazil—the fifth most populous nation in the world,” he said. “Facebook is estimated to count approximately 1.1 billion users, and is on track to overtake China and India and become the world’s most populous community.” At the same time, millions of images are uploaded daily in Flickr and Instagram, and over 70 hours of video are uploaded in YouTube every minute. Social media has undoubtedly changed the world, and its impact has also made it a powerful media for intelligence work. While people generally consider the Arab Spring events of 2011 as the watershed moment for social media, its impact was actually being felt much earlier. Twitter, Facebook and YouTube were used extensively in June 2009 to provide real-time accounts of the situation in the streets of Tehran during Iran’s Green Revolution. “That use of social media in Iran in 2009 was not planned, but rather something that just happened spontaneously, a revolution going viral,” Stefanidis commented. “That unrest eventually fizzled out, but it served notice to the Arab world, making them fully aware of the power of social media to bypass state-controlled news channels in order to bring their message to the attention of the rest of their countrymen, and to engage the world community.” A short 20 months after Tehran, social media was used in a concerted and organized manner in the Arab Spring.
Anthony Stefandis
Jeremy Balian
“Social media can help us understand political instability around different events, changes in government, or disasters and crises. This gives us the ability to tap into the pulse and see how they are reacting and experiencing.” No longer is analytics relegated to the periphery. “This social discourse is massive, and analytic technology must be applied to glean actionable intel from this very public conversation,” reported Jeremy Balian, director, social intelligence solutions for Topsy Labs. For example, he explained, the Arab Spring and other internationally significant social tipping points can be detected earlier by analyzing social media in real time. “This early warning capability ultimately helps to best prepare and respond to rapidly changing world events,” he said. Locked away in the big data of social media are little nuggets of information relevant to particular missions and activities that intel people are monitoring, noted George Demmy, chief technology officer of TerraGo Technologies. “An individual’s ability to realistically glean the relevant information and put it into some kind of context without the tools and supporting infrastructure is doubtful to say the least,” he said.
Human Landscape
The proliferation of social media is leading to the emergence of a new type of geospatial information. “We are no longer observing the terrain and its variations, but are focusing now on the human landscape,” observed Stefanidis. “We can George Demmy now observe groups of people, how they are formed, when they are formed, how they interact and how they respond to events around them. The information we collect is no longer just geospatial, but rather it is geosocial, and this requires the development of new solutions to make best use of it.” As a result, a number of firms are developing Mobile Revolution analytical software that provides the military and intelligence communities a suite of technologies for The rise of social media is distinct from, but closely discovering a cornucopia of content in social media related to, the booming popularity of mobile technology. and documents, including traditional geospatial The number of people with mobile devices is now outinformation as well. The business is highly competstripping those with desktops, noted Sean Gorman, chief Rick Cobb itive, with a number of companies coming out with strategist at the Esri research and development center in similar products that address different angles. Arlington, Va. “Behind that, they integrate the relevant information into intelli“That’s happening even at a great rate outside of the developed gence products that automate the forging stage and give the analyst world,” Gorman said. “What we’re seeing now is a whole bunch of peoa leg up on the sense making—or, once the sense making has been ple around the globe on mobile devices that are enabled with GPS to done, deliver intelligence products and applications to decision makers track what they are seeing and experiencing in real time.” and warfighters down range,” reported Demmy. The switch to mobile devices equipped with GPS is especially sigTerraGo offers a core set of technologies that include its GeoXray nificant because it adds location information, which enables visualizafamily of technologies. TerraGo Chief Executive Officer Rick Cobb tion and prediction of the spread of social movements, for example. described GeoXray as allowing users to discover “needles in the hayAs a result, in recent years, social media has become the center stack” by first analyzing what content belongs to a specific place, and of the Internet. For those in the military and intel community, these then going into what content belongs to a specific topic. “This then are interesting times indeed. “This is one of the most exciting forms of allows people to do temporal analysis,” he said. human geography we have seen in quite a while,” exclaimed Gorman. 22 | GIF 11.5
www.GIF-kmi.com
Essentially, GeoXray addresses the challenge of information overload by streamlining the process of discovering, monitoring and analyzing an overwhelming amount of big data from a wide variety of sources, including news feeds, blog posts, social media, maps and imagery, and more. It allows analysts to quickly filter results by place, time and topic to provide enhanced situational awareness, support better decision making, capitalize on opportunities and anticipate and avoid threats. “GeoXray allows users to define very specifically what a place or area of interest means to them,” Cobb said. GeoXray also allows people to connect to any source of information—including open source, social media, Facebook and Twitter, as well as internal sources. “GeoXray allows people to connect to any of these social media sites as well as internal sites of unstructured data and start to describe various topics in which they are interested,” Cobb said. “It’s not just keyword search. We combine groups of keywords to develop themes. We filter through that content that has, by the way, already been filtered by place.” The last thing is to filter by time. “When you get the combination of place, topic and time, and you are allowed to describe your location very discreetly and pick your sources, the amount of relevant data you end up with is extremely high. What we have done is create a sophisticated analytical approach based on open source intelligence, then pushed it out to the edge user to strengthen the analytical material available to an organization by magnitudes,” he said. “We see the industry moving from human geography as a static phenomena that we study through surveys and random sampling techniques, to one where we take our methodology and update it and evolve it with real-time data feeds that are coming with locations and unstructured date,” said Gorman. Esri then puts data in a format and runs an analysis across it so that it can be digested, thus enabling analysts to understand what is occurring in the data and what biases there might be. “We put error bounds around that in the decision-making tool and distract the signal from the noise,” he continued.
Real-Time Data Esri products include Geoevents Server, which lets users integrate and exploit real-time data streams with the ArcGIS system, and Geoportal Server, a free, open source product that enables discovery and use of geospatial resources including datasets, rasters and Web services. “We are also working on adding methodology for social data,” Gorman added. What makes Esri’s products different from those offered by other companies, he suggested, is the fact that Esri takes a geographic perspective of social media. “We are also looking at other aspects of the data like network components or unstructured data. We are focused on helping people deal with and interconnect the geographic aspects of social media.” Thermopylae Science and Technology offers iSpatial, which the company describes as a powerful software framework that enables visualization of limitless information, intelligence fusion, and operations management in a geospatially aware environment. The iSpatial framework leverages and is closely coupled with Google’s Earth Enterprise product as a visualization platform. ISpatial can be integrated with virtually any existing data source and serves as an enterprise visualization service to expose an organization’s data in a geotemporal user defined environment. www.GIF-kmi.com
“It was designed to bring in mapping elements and mapping tools,” Clark said. “Because it is web based, it can be shared.” Topsy executives maintain their company offers the largest social media search and analytics engine available today. “Just like Google for documents on the Internet, Topsy offers the largest search engine for social media on the web,” said Balian. “Our fast-indexing and liveranking technologies process massive amounts of authored content from the world’s largest social networks in real time, keeping everything forever.” The products Topsy offers to government organizations are Topsy Pro Analytics Public Sector, Topsy API Services, and Topsy Data Services. Topsy Pro Analytics Public Sector empowers government agencies to understand and identify key social signals critical to their mission. “Users can identify real-time emerging trends, events, indicators and warnings,” Balian said. “The product features a web dashboard that allows real-time and historical analysis of the entire Twitter firehose and provides a full array of social metrics that include sentiment, geography, influence and more.” Topsy API Services enables agencies looking for specific social data to acquire it via an application programming interface. This data includes proprietary Topsy social metrics such as relevance and influence. Versions of this API focus on news, search and trending. Topsy Data Services allows agencies to perform their own analysis by pulling large amounts of raw social data from Topsy. This data includes all publicly available Tweet content over multiple years, along with proprietary measurements that quantify the volume and characteristics of communication for any keyword, term or domain of interest. Another company that is active in this area is Lockheed Martin, which offers LM Wisdom, a predicative analytics and big data technology tool that monitors and analyzes rapidly changing open source intelligence data, including newspaper feeds and social media content.
Promise or Panacea The advantage to open source data is its size and scope. “We are able to get information in real time from huge numbers of participants across the globe in a way we were never able to reach by classified means,” remarked Gorman. Another obvious advantage is the tremendous opportunity it provides to expand data collection capabilities. “We can now collect data from the ground, even from denied areas,” said Stefanidis. “We can have this information available instantaneously as well. Suddenly we are presented with a hybrid massive sensor network that we can tap into, complementing our authoritative data sources.” Social media offers many promises, but is not a panacea. That truth was exemplified by the Tsarnaev brothers, who allegedly devised and carried out the recent Boston Marathon bombings but apparently left no evidence on the Internet about their intentions. “People expect social media to be some kind of magic bullet—that all things can be discovered by analysis of social media,” commented Cobb. “Frankly, that is not true. There is almost an equal amount of misinformation through social media as usable information.” Cobb pointed out how in Boston, the Tsarnaev brothers were highly discreet and clandestine. “There was not a community discussion that there was going to be this atrocity,” he said. “But, on the other hand, there were a lot of home-generated pictures and video surveillance— not exactly social media—that helped to catch [Dzhokhar Tsarnaev].” GIF 11.5 | 23
“Noise” also presents issues. “There’s a real challenge around understanding veracity,” remarked Gorman. “There are many ways to interject false information to mislead an analysis.” Consequently, it’s critical to understand the bias in social media and the demographics through which the information is streaming. A variety of approaches make it possible to do this. One is a sample theory. “Instead of focusing on individual messages, focus on trends across a very large scope of information that is self-verifying,” Gorman suggested. “It is one thing to have one person saying there has been, say, an attack on a particular location. But if hundreds of people are simultaneously saying the same thing, and they are doing it spontaneously without having contact with each other previously, then we can feel confident that the event is true.” Mathematical techniques can also be used for finding false streams versus actual occurring information. “The trick is not to get focused on reading specific messages and looking at microscopic trends,” Gorman said. Stefanidis acknowledged that the signal-to-noise ratio characteristics of social media generally do not meet the standards the intel community has been used to. “Valuable information is often hidden under massive amounts of trivial content,” he said. During Hurricane Sandy, for example, people provided on-thespot reports about flooding in New Jersey through Twitter and Flickr. But some of those reports also contained the fortunately incorrect information that sharks were swimming in the streets. “We need to be cognizant of these particularities when processing social media content. Luckily, the massive public participation across multiple platforms supports the development of effective validation solutions,” Stefanidis said. Repressive governments also have tried to fight social media through technology, for example by practically unplugging the Internet in their countries. “The Syrian government tried that in November 2012, and again in May 2013,” Stefanidis pointed out. “These efforts were proven futile, because technology allows the locals to still access the Internet through other means, such as dial-up.”
But Stefanidis said he is more concerned by the reported efforts of Chinese officials to affect social media content through the use of paid commentators who try to dilute discussions critical of the government. “This may become a new type of government-sponsored cyberwarfare, where the objective is not the denial of service, which the objective of hackers today, but rather the denial of message,” he warned.
Paradigm Shift Looking ahead, Gorman forecasts that mobile devices will continue to proliferate, with more social services built on top of these platforms, especially since there is greater incentive to include locations in addition to metadata. “We will see increasing value and reliance in open source information coming off of these devices,” he said. “We are in a paradigm shift.” Social media analytics tools entering the market today are primarily designed for a variety of end-user types, from the intelligence analyst to on-the-ground personnel, Balian noted. “These tools currently take a one-size-fits-all approach and do not yet allow the sort of customization required to effectively support mission needs.” But Balian sees users increasingly demanding tailored solutions, including social media analytics capabilities within their current tools. “Systems integrators will need to build focused applications for specific mission needs,” he said. “There will only be a few social media analytic providers delivering not just data feeds, but the core index, platform and measurements necessary for effective applications.” The strength of the underlying social media analytics foundation must be its awareness of the entire social dialogue, he added. “This complete picture of what folks are saying is a core requirement. Having or not having the combined real-time and historical perspective of the social conversation will ultimately determine the success or failure in applying these solutions to the community’s mission priorities.” O For more information, contact GIF Editor Harrison Donnelly at harrisond@kmimediagroup.com or search our online archives for related stories at www.gif-kmi.com.
R U O TH
12
The Navy’s shift to the Pacific inspires our twelfth title and website...
E TL TI
OUR INAUGURAL ISSUE
will support the Navy with the latest program developments in air and sea for Congress, the executive branch, other services and industry.
Contact Nikki James at nikkij@kmimediagroup.com or 301-670-5700 to participate in the inaugural issue!
24 | GIF 11.5
www.GIF-kmi.com
Going beyond the “Zero Dark Thirty” scenario, data triage can provide the key to understanding massive amounts of newfound information.
By Samuel L. Park and Chris A. Mattmann and the world wide web. A city of Sherlocks would The Navy SEALs proceeded with their mission, be needed just to eyeball the data, much less to draw with obvious urgency, floor by floor, sack by sack, gathclever inferences from it. Welcome to today’s world ering potentially valuable material along the way, affixof big data. ing information stickers and stacking it all for team Some quick assessment of the whole situation is extract. needed, and for big data this ultimately comes down As portrayed in the movie Zero Dark Thirty, SEAL to having a smart way to do data characterization at Team Six had just flawlessly executed the raid on ever finer levels. This data triage, important in itself, Osama bin Laden’s Abbottabad residence. But the misfurther provides a quick first characterization and sion was not over—not in a site worthy of Sensitive Site initial disposition of the newfound data. It thereby Exploitation attention. They also had to emerge with facilitates other exploitation of opportunistically disthe trove of potential information that could be there. Samuel L. Park covered big data. They egressed beyond the building perimeter, not restNew tools such as Apache Object Oriented Data ing until unburdening themselves at a rendezvous and Technology (OODT) and Apache Tika make this data debriefing area. triage more rapid, effective and easy to implement, While maintaining chain of custody, the unloaded, ultimately changing a practically infeasible challabeled material was transitioned to responsible analenge into something matter-of-course. OODT and lysts, who then faced their own daunting challenges. related capabilities can be quickly brought to bear, What kind of data was there? What was it? How much in combination with legacy and developing analytof it was there? Was it connected to other data on the ics. The resulting system can be ramped up quickly web or elsewhere? Most of all, what did it mean? to operational scope and scale in order to address an This Zero Dark Thirty scenario motivates considurgent problem, all the while maintaining key linkeration of the technical and operational challenges Chris A. Mattmann ages with subject matter experts. that face time-constrained intelligence extraction from New programs sponsored by the Defense newfound data troves of unknown size, characterisAdvanced Research Projects Agency (DARPA), especially the XDATA tics or content. Analysts must deal with voluminous multi-terabyte initiative, support these efforts and provide new capabilities for govsized hard drives and gigabyte sized cellphones, each likely containernment, non-profits and commercial enterprises. ing links to a much larger world of legacy databases, social media www.GIF-kmi.com
GIF 11.5 | 25
Big-Data Triage Big data is big in at least three ways—in volume, velocity and variety, sometimes called the “3 Vs” of big data. The initial challenge with a new cache of unknown data is its potential variety. It could be virtually anything that bits can be made to represent. Beyond the data’s size, its structure/architecture, volume and connectedness provide the first meaningful characterizations of it. The sheer number and continuing proliferation of file formats is by itself a significant challenge, even without considering encryption. According to FileExt, the file extension information website (http://fileext.com), there are more than 50,000 file types currently in use, and growing. Many of them require specialized, often proprietary, software to process and understand. The challenge becomes vastly more complicated as links to other data are discovered, which themselves can be linked to yet more. Sophisticated graphical models are needed just to capture the general overall structure. Further, in many sensor and social media applications, the large volume of data arrives in continual repetition. This velocity requires creating special storage and/or filtering/synthesizing of the data. At the start of this ever-expanding domain is the challenge of “data triage”: How does one go swiftly from the first characterizations to subsequent ones? NASA’s Jet Propulsion Laboratory (JPL) has been a pioneer in developing computational approaches that can handle such situations. OODT, a set of related collection and archiving tools, is one such innovation. OODT has been proven over the last decade for its use on a number of earth observing, climate and science missions. It was transitioned by NASA to the Apache Software Foundation (ASF) in 2010, the first NASA software project to do so. Apache OODT became a top-level ASF project in January 2011, and continues to rapidly develop new capability, specifically for applications in big science and in medicine. OODT provides tools that quickly wrap legacy or new applications, enabling them to be connected to each other, to data sources and to end-users with a minimum of fuss and expenditure. Apache OODT was recently selected as part of a proposal by MDA Information Systems (MDA), teamed with JPL, as a big data technology in DARPA’s new XDATA initiative. Some tools associated with OODT are ideal for “automated metadata extraction,” which essentially provides the first characterization of newfound data, including automated file-type detection over an extremely broad range of file types. The Apache Tika framework delivers this capability seamlessly. This was exercised as an example application during the XDATA program kickoff in January, where a set of example big data was “crawled,” characterized and interfaced with an iPhone. The results of this exercise are applicable to a number of domains, including big science, medicine and geospatial intelligence. For example, usage of geolocation data in social media mining and analytics presents big-data challenges, which can now be mitigated using OODT- and Tika-based data acquisition, metadata extraction, crawling and MIME type detection. Although OODT and Tika are admirably adaptable, they exist in a universe of software having many capable big-data alternative analytics, many of them having higher performance in niche processing roles. The special role of OODT is in taking the first whack at things, figuring out what data is there, wrapping legacy analytics, connecting all parties together and expanding to scale. After this is 26 | GIF 11.5
done, more specialized applications may be brought to bear. This is essentially the role called for in data triage. The wrapping of analytics is not only quick, but also effective by several measures. By using the legacy code itself, in an unaltered form that has been intimately vetted by subject matter experts (SMEs), it reduces errors that often arise from recoding. These errors are often extremely difficult to ferret out and can result in costly erroneous results and the follow-on consequences of them being used. OODT’s modularity also allows each SME’s analytics to be included in a parallel development process, further reducing calendar time to getting a result. Data triage comes in at least two parts: first, the quick characterization of data, then, the disposition of that data to other analytics. The second step calls for invoking workflows to accomplish these tasks. For big data, because of its size, such subsequent workflows often take some time to execute, though preliminary results are a common and quite useful characteristic as well. The coordination of these analytic steps is greatly facilitated by new visualization approaches. DARPA’s XDATA program has anticipated these needs. Of the 24 XDATA performers, eight are doing visualization development. MDA and JPL are working with these visualization performers on providing new, more approachable ways of taking advantage of these big-data analytics—ways that do not require command line expertise to function.
Free and Open Software Lastly, the OODT and Tika software themselves are free and open, meaning there is open knowledge of what is in the software and how it functions, there is an extended community of contributors and users, and the body of software itself is a ready source of “capital” for forward-looking business, non-profit organizations and government. This openly available new capital is a resource that all stakeholders can use to advantage or, conversely, ignore to their disadvantage. Answering the data triage challenge is the start to answering a whole lot more about big data. Leveraging its proven pedigree in big science, new tools associated with Apache OODT can be brought to bear in the geospatial community. OODT and related analytics can be brought to bear adaptively, scaled to operational scope, all the while using a wrapping approach that maintains maximum linkage with experts in the field. What is your geospatial big-data challenge? Perhaps you need to triage an abundantly endless collection of geo-tagged social media feeds? Or quickly assess information from your UAV and satellite imagery feeds? When faced with your next big-data opportunity, utilize these new open source tools that can make practical use of this data in a hurry: Link it in, scale it up and use it to the max.O Samuel L. Park is with MDA Information Systems. Chris A. Mattmann is with the Jet Propulsion Laboratory, California Institute of Technology, and Computer Science Department, University of Southern California. For more information, contact GIF Editor Harrison Donnelly at harrisond@kmimediagroup.com or search our online archives for related stories at www.gif-kmi.com.
www.GIF-kmi.com
The advertisers index is provided as a service to our readers. KMI cannot be held responsible for discrepancies due to last-minute changes or alterations.
GIF RESOURCE CENTER
Compiled by KMI Media Group staff
Advertisers Index
Calender
Actionable Intelligence for Defense and Homeland Security Summit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 www.actionableintelsummit.com BAE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 www.baesystems.com/gxp DigitalGlobe. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 www.digitalglobe.com /egd DynCorp. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C2 www.dyn-intl.com IHS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C4 www.ihs.com/geospatial Intelsat. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 www.spacedelivered.com TerraGo Technologies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 www.terragotech.com
July 8-12, 2013 Esri International User Conference San Diego, Calif. www.esri.com September 16-18, 2013 Air and Space Conference National Harbor, Md. www.afa.org
October 13-16, 2013 GEOINT 2013 Symposium Tampa, Fla. http://geoint2013.com October 29, 2013 SAP NS2 Solutions Summit Falls Church, Va. www.sapns2.com
?
Want to REACH the decision-makers in the DEFENSE COMMUNITY With a unique concentration on senior military officers and DoD leadership, KMI
KMI’S FAMILY OF PUBLICATIONS Border Threat Prevention and CBRNE Response
Media Group focuses on distinct and essential communities within the defense
GEOSPATIAL INTELLIGENCE FORUM
www.BCD-kmi.com
June 2012 Volume 1, Issue 1
Michael J. Fisher Chief U.S. Border Patrol U.S. Customs and Border Protection
precise way to reach the exact audience
MILITARY ADVANCED EDUCATION
that procures and deploys your systems,
MILITARY LOGISTICS FORUM
Robert S. Bray Assistant Administrator for Law Enforcement Director of the Federal Air Marshal Service Transportation Security Administration
MILITARY INFORMATION TECHNOLOGY MILITARY MEDICAL & VETERANS AFFAIRS FORUM G for Navy NAU I Medium
The Communication
The Communication
and most targeted distribution within critical
Leadership Insight:
Hazmat Disaster Response Wide Area Aerial Surveillance O Program Tactical Communications O P-3
GROUND COMBAT TECHNOLOGY
KMI Media Group offers by far the largest
Integrated Fixed Towers
Border Protector
market. This provides the most powerful and
services and equipment.
SPECIAL SECTION:
BORDER & CBRNE DEFENSE
SUE RAL IS PEOs U
Medium for Navy
PEOs
MILITARY TRAINING TECHNOLOGY
SPECIAL SECTION:
CARRIER ONBOARD DELIVERY OPTIONS
World’s Largest
market segments. Sharp editorial focus,
Carrier Craftsman
NAVY AIR/SEA PEO FORUM
Rear Adm. Thomas J. Moore
Distributed
Special Ops
Magazine
m www.npeo-kmi.co
U.S. Navy Program Executive Officer Aircraft Carriers 2013
SOCOM O
pinpoint accuracy and depth of circulation
PROGRAM MANAGEM ENT UPDATES
Craft Riverine Patrol Partnership
Shipboard Self-Defense Development Helicopter O Educational Presidential Munitions O Precision Guided
SPECIAL OPERATIONS TECHNOLOGY
SOF Enhancer Adm. Bill H. McRa ven Comman der Special Operatio ns Comman d
www.SOTECH-k mi.com
May 2013
Volume
make KMI Media Group publications
TACTICAL ISR TECHNOLOGY
the most cost-effective way to ensure your
U.S. COAST GUARD FORUM
11, Issue
4
Rapidly Deployab Robotics le Technolog Networks O SOF y O Global Light SOF TrainingVehicles
advertising message has true impact. To learn about advertising opportunities, call KMI Media Group at 301.670.5700
www.GIF-kmi.com
GIF 11.5 | 27
INDUSTRY INTERVIEW
Geospatial Intelligence Forum
Richard M. Cobb President and Chief Executive Officer TerraGo Q: Tell us about the new TerraGo Vision platform. A: The TerraGo Vision platform represents years of innovation with thought-leading customers and partners solving next-generation intelligence challenges. Today’s global enterprises need location intelligence [LI] solutions that reach beyond centralized, internal data stores and GIS systems—to include information and web services from open and commercial sources—and facilitate the delivery of timely intelligence within and across the enterprise. With an explosion in enterprise, commercial and open source content from myriad publishers, sensors and intelligence providers, big data offers incredible opportunities and daunting challenges for organizations seeking enhanced situational awareness, faster response times and improved decision making. To enhance collaboration and mission effectiveness, information must be intelligently filtered by place, time and topic. From there, it needs to be seamlessly integrated then rapidly delivered to decision makers, analysts and field personnel. TerraGo addresses this need with the TerraGo Vision platform, which aids in placebased discovery and monitoring, as well as the automated integration of feature-rich maps, high-resolution imagery, unstructured data and dynamic services. Vision also enables geospatial visualization on industrystandard platforms and enables the creation of deployable, collaborative, dynamic intelligence products on-demand. Q: How does this solution support the enterprise? A: The platform provides application services, developer tools and client/mobile applications in a system architecture that serves as a scalable and extensible foundation for mission-specific, on-demand LI applications. It easily adapts to real-life operational scenarios and fits modularly into established enterprise workflows. For example, enterprises can leverage their existing GIS, remote sensing and IT investments and add TerraGo 28 | GIF 11.5
LI and geospatial collaboration capabilities. Software OEMs can integrate TerraGo functionality into software products, while system integrators can rapidly design their own solutions for unique customer requirements. Finally, intelligence providers can introduce new offerings and expand their existing portfolios to reach new users with new products and capabilities. TerraGo Vision is proven in each of the above scenarios. Q: Tell us about the market need for this kind of solution. A: The need for LI discovery, integration and delivery spans a growing number of industries and operations. Whether in planning and executing military operations, responding to crises, protecting personnel and infrastructure, or monitoring and reporting customer and competitor activity, TerraGo Vision offers valuable capabilities essential to a comprehensive LI solution. The platform allows enterprises to rapidly access large volumes of data—from maps, imagery and documents to newsfeeds, blogs, social media and web services—and deliver timely, relevant location intelligence to anyone, anywhere operating in offline and connected environments. Q: How does this offering advance TerraGo’s evolution? A: TerraGo has a rich history of delivering GeoPDF solutions that empower organizations to produce on-demand intelligence products and to access, update and share portable, interactive, dynamic GEOINT.
Today, partly through the acquisition of fellow In-Q-Tel company Geosemble, we have evolved into a firm that also helps organizations automate the process of discovering, analyzing and geospatially visualizing relevant unstructured information from any source. We now address a broader spectrum of the most complex, pressing LI challenges. We know that a holistic LI solution requires many different technologies and services, so TerraGo Vision has an open architecture, designed to be partner friendly. Vision allows us to provide our technology as part of a custom solution or standard application, across desktop, mobile or cloud environments. Q: How do you see the role of GEOINT changing in the future for the defense and intelligence communities? A: We are in a thoroughly challenging time and the need for GEOINT is greater than ever. National security threats and natural disasters are on the rise, amplifying the need for quality, actionable GEOINT. At the same time, resources are on the decline. Defense and intelligence organizations must glean intelligence from massive amounts of data. This data is generated and morphs at an astounding pace, making the challenge of discovering information that is relevant to the mission, and putting it into spatial context, even more difficult. Delivering the derived intelligence to the entire enterprise in a timely fashion is essential to mission success. Thankfully, new solutions can help analysts automate much of the intelligencegathering process, leveraging technologies that continuously monitor a vast array of data sources, putting information in context of place, time and topic. Furthermore, automated approaches exist to integrating, visualizing and delivering geospatial intelligence to decision makers across the enterprise. These types of automated, collaborative geospatial intelligence solutions will play an increasingly larger role aiding critical decision making in support of military missions, crisis response and humanitarian efforts. O www.GIF-kmi.com
NEXT ISSUE
September 2013 Volume 11, Issue 6
Cover and In-Depth Interview with:
Gen. Keith B. Alexander Commander, U.S. Cyber Command Director, National Security Agency Chief, Central Security Service (Invited)
Features: Multi-INT Systems
Industry and the military and intelligence communities are developing new ways to bring streams of data from different sources together for integrated analysis.
Full Motion Video
As full motion video becomes an increasingly important element in the intelligence toolbox, the search is on for technology that will not only help in processing and exploiting FMV data, but also make it better.
Location-Based Cybersecurity
While effective cybersecurity is vital for protecting geospatial and intelligence networks, location-based information is also proving extremely useful in tracking down and defeating cyber-threats.
GEOINT Online
Geospatial application programming interfaces (API), such as the Maps Engine API recently released by Google, enable developers to easily integrate maps and other capabilities into their projects.
Special Report
GEOINT 2013 Symposium Preview
Insertion Order Deadline: August 9, 2013 • Ad Materials Deadline: August 16, 2013