Intelligence Beyond Imagery
INTELLIGENCE BEYOND IMAGERY
The National Geospatial-Intelligence Agency is on the cusp of dramatic
advances in GEOINT, as it focuses on using new technology, sensors, tools
and methods to support advanced geospatial intelligence.
By Marty Kauchak
The National Geospatial-Intelligence Agency is on the cusp of dramatic
advances in GEOINT, as it focuses on using new technology, sensors, tools
and methods to support advanced geospatial intelligence.
By Marty Kauchak
The National Geospatial-Intelligence Agency (NGA) is on the cusp of dramatic advances in GEOINT, as its focus on using new technology, sensors, tools and methods to support Advanced Geospatial Intelligence (AGI) paves the way for National Air and Space Intelligence Center (NASIC) and other community members to obtain a more complete perspective of its charts, maps and other pixel-based products, and the data behind those pictures.
The comparison between the community’s legacy geospatial intelligence and AGI is an issue of literal versus non-literal exploitation of GEOINT products, according to Frank Avila, technical executive with the NGA Office of Sciences and Methodologies.
“In traditional imagery analysis, the practices that we have had for the past several decades incorporate literal analysis of imagery data. You are trying to discern and to analyze the visual information of the spatial content that is captured in the image that you are using. You rely on using the human eye and the brain as your main processor,” Avila said.
AGI, by contrast, uses non-literal aspects of the metadata found within pixel information derived from the community’s sensors (commercial, civil, national, airborne, open source and satellite) that cover the electromagnetic spectrum to improve or enhance GEOINT analysis.
“This information can be derived in the form of technical, geospatial and intelligence information that is primarily derived by using advanced image-processing algorithms. There is data fusion that goes on and other techniques that we apply to really discern the non-literal content in the imagery data,” he noted.
As an example of the opportunities to support the end user with AGI, Avila cited the case of an analyst looking at a facility for visual cues that would confirm an event such as a chemical spill might have taken place. “Say we have some tip-off that a chemical spill took place at a specific facility. Although we might not be able to discern that, in fact, a chemical was spilled, there might be other observables that we can go after that might lead us in that direction, to make that final assessment,” said Avila, noting that one observable might be stressed vegetation gleaned through multi-spectral sources.
“It’s not going after that specific item we want to find out, but what are some other clues that will point us in the direction that, ‘Yes, that is what we think happened and here are the reasons why, and here are the observables that we were able to piece together or discern to lead us to that final conclusion with some level of confidence,’” he added.
A contemporary case of how AGI works was NGA’s efforts in support of Hurricane Katrina relief and recovery operations. The agency provided support to detect contaminated areas and determine where chemicals or oil had been spilled. In some cases people were advised not to go into potentially dangerous areas.
But in an era where a picture is more than ever worth a 1,000 words, the intelligence community still needs to help the end user visualize what data he or she is getting through non-literal, advanced processing techniques.
“One of our big challenges is how do we help visualize that into some sort of product so the end user can decide what he or she is going to do, and gain their trust and confidence because you are showing them something that they cannot see,” said Avila. “They have to be able to trust the information, and we, as the imagery scientists, have to be able to explain how the information was derived, the confidence level that we have on the data that was used to derive that information, and to portray that picture and help them visualize, ‘What is this really telling us?’”
The AGI community frequently describes usable data that is derived from the “phenomenology” of the science data it receives. This analysis requires subject matter experts from geology, botany and other earth sciences, in addition to engineering and other disciplines, who share the bond of having a common remote sensing background to describe changes in natural processes.
“They help answer the question, ‘How does that target, piece of soil, vegetation or camouflage behave in different portions of the spectrum?’ We really need subject matter experts in some of the different sciences, geology for example, that can help us understand when we looking at minerals or rocks, ‘What can we say about that target as we see it through various lenses across that spectrum?’” offered Avila.
MASINT Roots
AGI during its formative period was classified as part of measurement and signature intelligence (MASINT). But while many intelligence community members still refer to MASINT and AGI in the same breath, there are differences between the two disciplines.
“MASINT exploitation incorporates the use of signatures to interpret data and to characterize and identify targets of interest,” said Rob Young, a spokesman for NASIC. Phenomena MASINT collectors include electro-optical, radar, polarimetric and radio-frequency emanations, he noted.
“These overlap many of the phenomena used in AGI sensor employment and data exploitation,” Young said. “As such, they constitute valuable corollary data. When integrated using innovative methodology, the intelligence analysis outcome is greater than it would have been using either data source in isolation.”
Avila emphasized that MASINT and all other available information is used during analysis. “As we try to do our geospatial intelligence analysis, we don’t want to rely on a single source of information. We are trying to paint the best picture of the environment possible for the warfighter or the decision-maker,” he said.
“We use any piece of information that helps us paint that better picture as complete as possible,” he added. “MASINT has a lot of good attributes that we can fold into the work we do, just like SIGINT and HUMINT. It’s not an issue of whether we use MASINT or AGI, but which one plays which role in the issue at hand that we are trying to analyze.”
NGA’s organizational structure has evolved during the last two years to expand the use of AGI across the agency enterprise. In 2007 the functions of a dedicated AGI office were transferred into the agency’s analysis and production elements.
“One element is the current operations organization, which integrated the imagery scientists and analytical methodologists to ensure we can do full-spectrum exploitation of our remote sensing capabilities,” Avila said.
A second element of the former office of AGI, dedicated to technical and imagery science support, evolved into its own entity—Avila’s office. “The creation of this office marks an important milestone in our imagery science history, because now the scientific contribution and dimension is valued equally, at the same level, as our traditional imagery and mapping component,” he said.
Industry Interface
NGA’s collaborative efforts with industry and the academic community have potential to help solve evolving AGI technology gaps.
NextView is an NGA program that acquires satellite and remotely sensed imagery from commercial providers. The volume of data from imagery received from the NextView satellites presents opportunities, including automating data processing, for the industry-academic team.
“We don’t have enough eyes to look at all of this data,” Avila acknowledged. “One of the things we need are better tools to, in some cases, automate the processes of analyzing and processing some of these data sets, perhaps even before it gets to an analyst’s desktop. There are a number of things that could be done to research an image and cue the end user whether the activity or target of interest is found in that image so the analyst would not have to spend time analyzing that image, or perhaps learn that he or she wouldn’t have to analyze an image.”
An increase in imagery volume also provides new data storage, archiving and dissemination requirements. “When you talk about the volumes, in some cases gigabytes and terabytes a day, of data that we are storing and pushing around, this could produce a monumental problem. So we are looking for better ways to archive and store and disseminate the data,” he noted.
At the end of the dissemination process, Avila continued, the NGA team is also attempting to answer, “What’s the real product that should come out of AGI processing so we get to the end user or warfighter in the field who has the ‘soda straw’ for a communication line to receive information they need in a timely fashion?
“We should not send too much or too little, but enough so they feel comfortable with what they get and so they can incorporate that with other sources at their station and still be able to make some sense out of it and take required action,” he said.
Asked to list NASIC’s top three requirements to more efficiently collect, analyze and distribute AGI/MASINT products, Young responded this way: “The question is the stuff entire studies are based upon. There may be no ‘right’ answer. Rather than attempt it, suffice to say that critical requirements exist in each of the three areas mentioned.”
Hyperspectral technology, for example, has been studied and developed extensively, but development and deployment of systems that take advantage has lagged. “Algorithm development and overall innovative technique application have greatly enhanced analysis capabilities; integration that permits the timely fusion of multiple inputs from unlike sensors and sources is still in its infancy. Millions have been spent on IT solutions to dissemination, but users are still daunted by the perplexity of the multiple portal capabilities fielded to serve as their data reference points,” said Young.
Commercial Tools
AGI incorporates a number of commercial tools and services across the GEOINT portfolio, including for MASINT.
Leica Geosystems Geospatial Imaging’s AGI/MASINT support is an extension of embedded geospatial intelligence, enabling customers to benefit from the company’s client side and server-side products.
“Leica Geosystems focuses its efforts on image-oriented technologies—specifically electro-optical (EO), radar hyperspectral, multispectral, LIDAR (Light Detection and Ranging) and the ability to manage intelligence from the intersection of these various data types,” remarked Chris Tucker, senior vice president, Americas and national programs, Leica Geosystems.
The company is completing a number of projects in these areas. “Leveraging Erdas Imagine, Leica Geosystems will soon provide its customers with breakthrough capabilities for extracting intelligence from AGI/MASINT sources. Later this year, the company will introduce a dynamic new object-based feature extraction tool,” Tucker indicated.
The company is also participating in the development of unmanned aerial system (UAS) ISR sensor suites, which include the entire real-time processing and management chain. “While the weight and power limitations of the smaller UAS platform focus much of this work in the EO world, Leica Geosystems has addressed some very challenging problems with the integration of airborne EO, spectral thermal and even LIDAR on larger platforms. Leica Geosystems’ solutions focus on collapsing the processing chain of these data sources, and enabling warfighters with near-real-time exploitation-ready data from these sources,” said Tucker.
Analytical Graphics Inc.’s (AGI) commercial-off-the-shelf suite of software analysis tools allows its customers supporting AGI and MASINT missions to plan, collect and analyze collections in a pre-mission, real-time, and-or post-mission environment. “AGI has provided technology for the AGI/MASINT tradecraft since the days of the Central MASINT Organization. AGI products fuse geodynamic information from multiple sources and sensors, and present the information in a 4-D display,” explained Travis Langster, AGI business development director, intelligence community, and Tim Craychee, AGI systems engineer.
Reflecting on the evolution of the company’s products, Langster and Craychee observed, “The most significant difference is how AGI’s software is delivered. Traditionally we have served desktop users. Now, our software architecture delivers ‘software in any form.’ This approach supports traditional desktop and engineering analysis, the development of mission-specific applications using our software engine, and the use of software components for enterprise integration/service-oriented architectures. This format permits the GEOINT community to harness the value of more than three millions lines of code in a COTS technology that is available today.”
Ball Aerospace, meanwhile, has a 25-year engineering services history with NASIC, providing various corporate competencies.
“Ball scientists and engineers are national experts in processing and exploitation of radar, electro-optics, spectral and infrared data,” remarked Richard Juarez, senior program manager, Ball Aerospace. “We are able to provide support in processing and exploitation of MASINT remote sensor data because we specialize in developing state-of-the-art computer workbenches that host data processing algorithms.
“These workbenches seamlessly integrate mathematical algorithms, data fusion tools, graphical interfaces, report generation utilities and database libraries to provide analysts with user-friendly data exploitation solutions,” noted Juarez.
One significant opportunity on industry’s horizon is the request for proposals and contract award that are expected later this year for the NASIC umbrella contract for MASINT/AGI. Known as the Advanced Technical Exploitation Program, the contract will determine the level and quality of industry support to developmental and operational initiatives. “As you can imagine, this is a very sensitive procurement action. In fact, it is under the direct oversight of the Secretary of the Air Force Procurement Executive Office for Combat Mission Support,” Young said.
In October 2007, NGC and Science Applications International Corp. announced the formation of a team to bid on the program. Ball Aerospace is also competing for the contract. ♦
