Pacific Geomatics (PacGeo) is going to be offering AI-based clarification and refinement of satellite imagery through a new partnership for SuperX.
The satellite imagery provider will be partnering with Korea’s SI Analytics (SIA), a provider of AI-driven Earth observation solutions, to bring SIA’s SuperX AI-based image processing tool to customers as part of PacGeo’s imagery procurement and enhancement services.
Specifically, PacGeo said that the new arrangement is a strategic partnership with SIA, a “leader in AI-driven Earth observation solutions,” and that the partnership will “expand the global reach of SuperX.”
SuperX is a “generative AI-powered satellite image super-resolution technology,” according to PacGeo’s announcement. It is “built upon SIA’s proprietary Foundation Models” and can “process and learn from multi-resolution satellite data.” By incorporating this satellite imagery as training data, which they say can range from “0.25m ultra-high-resolution imagery to low-resolution global datasets,” the release said that SuperX is capable of “doubling the effective resolution” of satellite imagery.
SpaceQ reached out to PacGeo, and they provided further information.
Satellite imagery procurement
PacGeo is a satellite imagery reseller with a focus on procuring and providing satellite imagery, and the spokesperson said that they offer “vendor-agnostic expertise” to help clients find appropriate sources of satellite imagery. Their site says that PacGeo offers clients access to both “up-to-date and historical images” from “the most advanced satellites in space,” as well as “expert guidance” to help clients discover what imagery they need and how to glean insights from it.
PacGeo doesn’t operate satellites themselves nor provide analytic services, but have partnerships with a number of different imagery providers: Airbus, Vantor (formerly Maxar Intelligence), Planet, Capella Space, Satellogic, Pixxel, and Canada’s Wyvern among others. The imagery can include traditional optical data, but can also include multispectral/hyperspectral imagery, as well as synthetic aperture radar (SAR) satellite imagery and digital elevation data.
The spokesperson said that key client sectors include mining, forestry, national defense, energy, and intelligence. Notably, PacGeo represents multiple industry-leading satellite data vendors via National Master Standing Offers, including Vantor, Capella Space Satellogic, Wyvern, Pixxel, and SI Imaging Services — the sister company of SI Analytics. This gives DND and other Canadian Federal Government agencies a procurement path to access a variety of mission-ready datasets.
Applicable to fresh and archived imagery
As to why they decided to partner with SIA on SuperX, the model’s agnostic approach to imagery appears to be a key factor. The spokesperson pointed to how SuperX is “sensor-agnostic” — in other words, it works with a wide variety of imagery — and how that “aligns with PacGeo’s vendor-neutral model.”
(Notably, however, the spokesperson said that “SuperX is specifically designed to enhance optical satellite imagery.” Hyperspectral and multispectral imagery is already commonly analyzed using machine learning AI algorithms.)
The spokesperson pointed to how the technology is “applicable to fresh and archived imagery,” which opens up the opportunity to apply it to imagery that clients already have, instead of having to buy everything new. The clarity will be “useful for GEOINT (geospatial intelligence) analysis” — likely to be important to their national security clientele — as it will help analysts “interpret object shape, orientation, movement corridors, infrastructure details, and land-use patterns.”
The spokesperson also suggested a possible cost savings benefit, noting that it “improves effective resolution without requiring new satellite tasking,” which may help with “lowering acquisition costs and improving overall value.” In fact, in addition to offering SuperX as a part of image procurement, PacGeo will also be working with clients to apply SuperX to existing imagery.
In the release, SIA’s CEO Taegyun Jeon said that “by integrating generative AI with foundation models trained across multiple resolutions, we’re enabling machines not only to analyze imagery — but to understand, reconstruct, and predict what’s on Earth.”
Matt Tomlins, President of PacGeo, said that the partnership ensures that PacGeo has “Canadian exclusivity to one of the most advanced AI-powered super-resolution tools for fresh and archive satellite imagery”, and that the deal will help them “make high-resolution satellite imagery more accessible and cost-effective for users across industries.”