Real Technology, Real Benefits is a series of articles that explores the connection between satellite imaging technology and real-world applications and benefits. Read Part 1: pointing agility, Part 2: revisit rate & collection capacity and Part 3: image quality.
In the last blog post, we emphasized the need to use the right tool for the job, and implied that there is a role for satellites of varying capabilities. The premise is quite simple: while we believe there is limited value to our customers and partners from accessing low-resolution satellite imagery alone, there is much greater potential for unlocking value when these capabilities are combined—in a well-orchestrated manner—with DigitalGlobe’s high-resolution, high-accuracy, broad-area imagery.
Last week we took another big step toward at delivering this value to our customers with the announcement of our partnership with TAQNIA and KACST to develop, operate, and market a constellation of at least six highly capable small satellites. These satellites will collect imagery with 80 cm resolution, and they will leverage DigitalGlobe’s industry-leading ground network, imagery production expertise, content management, Geospatial Big Data Platform (GBDX), and global distribution channels to enhance a wide range of applications for customers around the world.
Including these small satellites, and following the deployment of our next-generation satellite architecture to replace WorldView-1 and WorldView-2, the DigitalGlobe constellation, with the incorporation of our Saudi partner’s capacity, will have the ability to revisit high-interest areas a stunning 40 or more times per day. So what are the real-world benefits to accessing a constellation optimized for high coverage, high resolution, high accuracy, and high revisit?
Today, our WorldView-class satellites image huge areas with the highest level of detail in the industry. We use a variety of techniques, such as crowdsourcing and automated feature extraction (based on recent advances in deep learning), to identify areas and objects of interest. For example, we recently leveraged our imagery, our population density algorithm, and our Tomnod crowd to map every human settlement in more than 333,000 sq. km. in northern Somalia and parts of Ethiopia with nearly 100% precision to allow for the most efficient delivery of polio vaccines. GBDX allows these types of sophisticated analysis at scale by allowing an ecosystem of algorithms and image processing routines to run at blazing speed next to the data.
Pan and zoom around this map to see the locations where DigitalGlobe’s algorithm detected human settlements.
Bringing the small satellite constellation into service in late 2018 or early 2019 opens up new opportunities. We will be able to detect change in known areas on a frequent basis with our smaller satellites, and then zoom in to understand that change with great fidelity using our high-resolution satellites. Conversely, when we identify a previously unknown area of interest with our high-resolution satellites, we will task our small satellites to monitor that area for change — up to 40 times a day. Key to making this work is our ability to cross-cue between satellites by leveraging our advanced tasking system that is capable of managing priorities across a multi-source constellation. In addition, we will uniquely be able to improve the positional accuracy of small satellite imagery by referencing it to our global high-accuracy imagery, much like some wrist watches use radio signals to match their time to extremely precise atomic clocks.
In the notional example below, WorldView-3 images the coastline of this small island, showing an empty beach (figure 1, 30 cm GSD); then, a small satellite flying directly over the island captures an image, showing no change (figure 2, 80 cm GSD); hours later, another small satellite takes an off-nadir image of the area, detecting a new presence on the beach. (figure 3, 1.3 m GSD); this cues WorldView-2 to take a closer look, showing military vehicles and equipment that an analyst could identify. (figure 4, 50 cm GSD)
Click to enlarge.
But the promise of a multi-source environment won’t have to wait until 2019, as DigitalGlobe already offers a range of third-party data sources through various platforms. We offer multiple commercial and open data sources to our oil & gas customers via the Spatial On Demand platform. And we have an agreement with SI Imaging Services of South Korea to offer access to Kompsat-2 and Kompsat-3 imagery to enable us to satisfy strategic customer demand faster in capacity-constrained regions where we are allocating our 30 cm capacity to other tasks. These arrangements are valuable for our partners, enabling them to reach a broader set of customers by leveraging DigitalGlobe’s global sales channel and addressing use cases enabled by the combined DigitalGlobe and partner constellations.
We are not limiting our partnerships to the optical imaging realm, having engaged with radar data providers to help bring the cloud-defeating and dawn/dusk observation capabilities that SAR platforms offer. Our partnership with MacDonald, Dettwiler and Associates to offer RADARSAT-2 satellite data enables yet additional use cases, like finding ships by imaging through the clouds with radar to cue our high-resolution satellites for a closer look.
Finally, our customers are also keen to explore the value that free and open data sources can provide to their business and mission needs. As civil governments make more data streams available to the public, we fully intend to help unlock the value they bring to the table. Landsat 8 and its follow-on missions, the European Copernicus missions, and many others across the globe promise to bring more interesting opportunities to our customers and partners, such as identification of broad area urban change or deforestation.
These new data sources benefit users that are interested in having as much reliable information as possible about a particular area in order to make the most informed decisions possible. These data sources also fuel a different class of applications, those that seek to understand where certain conditions exist across very large areas, and in some cases, across spans of time. These “show me where” questions are why we built our Geospatial Big Data platform (GBDX) – to answer questions such as, which commercial timber farms have expanded or contracted over the past year around the world, or which unconventional oil extraction sites are in active operation? The potential to unlock the value in our massive, multi-source imagery catalog has attracted many ecosystem partners and customers in the past year, from start-up algorithm and application developers like Orbital Insight and Precision Hawk, to blue-chip firms like Harris and Facebook. As our ecosystem grows, we are giving developers and customers access to a growing catalog of historical and current data and the tools to make the most of it.
DigitalGlobe has always been at the cutting edge of our industry. We are now extending our global leadership position by re-inventing the art of the possible with integrated multi-source capabilities, cloud computing and machine learning technologies, and Geospatial Big Data. It’s easy to see why more customers, data providers, and algorithm and application developers are choosing to join our ecosystem as together we build The Digital Globe.
Dr. Walter Scott is DigitalGlobe’s Founder and Chief Technical Officer.