Tuesday’s session, From Space to Mud and Beyond, featured four esteemed panelists representing GEOINT supply chain from sensor through data onward to decision: Colonel Kevin Golinghorst, commander and district engineer at the St. Louis District U.S. Army Corp of Engineers; Ann Carbonell, Ph.D., principal/senior advisor at Riverside Research; Joe Morrison, VP of Commercial Product at Umbra; and Gabe Chang, Federal CTO Architect at IBM.
On the topic of spatial cutting-edge innovation, Morrison mentioned that about five or six years ago, the industry was focused on building platforms for processing remote sensing data, but today, the innovation is happening on top of that infrastructure — what people are doing with those platforms and how they work together.
“The question is the platforms are built, so what do you do with them, how do you apply them?” he said. “I think the most interesting cutting edge work that’s happening starts with a problem and only uses a little bit of remote sensing to solve it … The ratio of remote sensing data to non-remote sensing data is maybe 10 or 20 to 1; the amount of other data sets that you have to bring together to be able to make a magical user experience where all the right data at just the right time is presented at just the right moment to make an informed decision and react to something really quickly — that’s not just a satellite/ML/AI problem, that’s a software, systems and process problem, too.”
On the topic of earth imaging and collaboration, Golinghorst touched on the U.S. Army Corp of Engineers’ fusion of satellite and both manned and unmanned platforms. “Here in St. Louis, we have a technical center of expertise, and we work with eight major contractors on a five-year contract to cover any customer across the nation and around the world from federal customers even down to the state and local level … Both in-house and leveraging contract capabilities is a key part of the Corp of Engineers’ mission.”
To close out the discussion, the panelists addressed the vague concept “data fabric” and how it connects space to mud and beyond. “The management of data fabric includes not just data at rest but also data in motion, the integration of that data and the virtualization piece,” Chang said. “It’s important to have a master data plan strategy, and on the security side, you have to take into account both outward and insider threats and enforcing policies across multiple clouds.”
That poses this question moving forward: With all this data coming in, is there still a need for synthetic data? Carbonell said yes. “There are a lot of problem sets where we have fleeting targets and sparse data … We may not see a capability, platform, launch or test, so when you don’t have those opportunities, somehow you have to create that synthetic data so that you can start training the machine-learning algorithms. And as we get data, we can replace that synthetic data.”