The Defense Advanced Research Projects Agency (DARPA) recently awarded Descartes Labs, a $2.9 million contract to bring its geospatial analytics platform into DARPA’s geospatial cloud.
Keen to understand how this contract will enable both Descartes Labs and DARPA to use geospatial for achieving more milestones in community development, we had an interaction with Mark Johnson, CEO, Descartes Labs and Jon Heinsius, General Manager & Director of Geospatial Programs, GeoNorth .
What’s the use of artificial intelligence in satellite imagery analysis and how is Descartes Labs pioneering this effort?
The Descartes Labs platform allows data scientists to perform real-time analysis of massive amounts of data. With the use of artificial intelligence, data scientists can focus on building out global-scale analysis and making these insights actionable, rather than spending time processing, curating and prepping data.
What does the collaboration with DARPA actually entail? Are you building this platform exclusively for DARPA or is it also commercially available for others?
DARPA has selected Descartes Labs to power its new Geospatial Cloud Analytics (GCA) program.
The significance of the GCA program is to demonstrate rapid global analytics, leveraging an array of commercial and other non-classified sources, all stored in the cloud. The community will benefit as a result of GCA creating a framework and fundamental tools (including a cloud-based data repository and analytics marketplace) which can be leveraged by analysts and analytic programs, allowing analysts/users to focus on their analytic challenges – rather than creating, managing, and maintaining the content they use for their analyses.
GCA will aggregate geospatial data into one massive, streamlined virtual data repository where government analysts can access the most up-to-date images anywhere in the world and advanced tools to analyze them. The Descartes Labs platform features a cloud-native infrastructure designed to provide the storage, computing, access, and tools needed to analyze massive, complex geospatial datasets, making it an ideal foundation for this DARPA program.
Under the GCA program, teams selected by DARPA will use the Descartes Labs Platform to build global-scale applications for issues including food security, fracking, and illegal fishing. The applications will be offered in GCA’s marketplace as a commercial service. To support these objectives, and pave the way for the development of additional applications, Descartes Labs will integrate up to 75 new datasets sourced from members of a diverse data partner network.
The GCA effort will roll out in phases over the course of two years. Phase 1 will orient analytics providers and prepare platform offerings for competitive comparisons. Phase 2 will be a one year competition between analytic solutions and platform offerings. Phase 3 is an optional exploration of different marketplace and economic models. Descartes Labs has been awarded $2.9M for phase 1 of the GCA project with a phase 2 award option of $4.2M.
The major benefit is that the Descartes Labs Platform simplifies many of the operations government analysts face on a daily basis and offers substantially more compute performance than is otherwise available.
Where will we see the immediate effect of this collaboration?
Through this collaboration, data scientists will be able to extract actionable insights from the vast amount of data available through the Descartes Labs Platform. GCA will develop an analysis-as-a-service marketplace from the existing data repository so that scientists can deliver intelligent applications that manage real-world issues such as food security and illegal fishing.
The data repository, when combined with the machine learning capabilities of the Descartes Labs Platform, will ensure that this group can focus on running analysis on a wide swath of cleaned-up, prepped data, rather than spend time gathering and processing raw satellite imagery.
What is the role of GeoNorth in the GCA program?
GeoNorth is the supplier of data and information that will be input to Descartes Labs and other winners. According to Jon, GeoNorth Information Systems’ (GNIS) primary role in the GCA program is to provision geospatial data to Descartes Labs for both the Data Platform and the Analytical Applications portions of the GCA program. GCA program objectives include intensive use of a broad range of geospatial data sources, in large volumes. GNIS offers expertise and experience in sourcing geospatial data from a wide range of data providers, including commercial, government and scientific sources.
GNIS is enabling Descartes Labs in their role as providers to the program, by providing the necessary expertise and services for provisioning source, geospatial data. This will enable Descartes Labs to focus on the higher-level requirements associated with the Data Platform and Analytic Applications of the GCA program.
In 2016, Descartes Labs was awarded a grant from DARPA as part of its SBIR Phase II program. This grant was the first example of Descartes Labs making sensor data useful by ingesting, scrubbing, and cleaning it up for faster use and analysis. The specific project had Descartes Labs analyze, monitor and forecast wheat crop across the Middle East and North Africa, with the goal of extracting correlations between crop health, and human changes, such as population growth and migration. The work helped expand our understanding of remote sensing data and allowed DARPA gain new insights into the availability of natural resources in the area and their impact on food supply, to get ahead of potential political unrest.
The recent award will bring more such impactful initiatives to fruition enabling us to have a more sustainable world.