When imagery from the Worldview Legion satellites revealed fabric over a large object near a dry dock in North Korea in September, satellite operator Maxar Intelligence knew they needed help. They turned to Umbra Space, a company that focuses on synthetic-aperture radar imagery, for follow up. Umbra’s sensors are not obstructed by clouds, darkness or even thin fabrics. And so after tasking a satellite on orbit, Umbra’s team ultimately revealed what was under the netting: the outline of a submarine.
Such intelligence provided a coup for North Korea observers and offered valuable insight into what the country’s military may be developing.
But the bigger feat may be in how such intelligence was gathered.
Coordinating observations of space-based sensors, known as tipping and cueing, has been an industry goal for decades. In spite of daunting technical challenges, tipping and cueing is now becoming more common and more seamless thanks in part to the proliferation of satellite sensors and artificial intelligence, which helps automate the task of sharing the geographic location of an interesting site or the trajectory of a moving vehicle or vessel. It’s also accelerating the fusion of data from complementary sensors.
“The real opportunity is making Earth-observation satellites work together to deliver actionable intelligence,” said Brian O’Toole, the CEO of imagery and analytics firm BlackSky. “Most customers don’t want to process raw data. They just want the answers.”
Solving the puzzle
Many of the tips that automatically prompt satellites to gather imagery or data come from news feeds and public data streams, such as U.S. Geological Survey earthquake alerts. Satellites also run algorithms onboard to process imagery in search of the unexpected.
That’s a tricky equation because “the part of that system generating tips needs to know what to look for,” James Mason, Planet chief space officer, told SpaceNews by email. “If it’s too generic/broad then it can make a lot of useless false positives. If it’s too narrow it could miss important events.”
What’s more, a satellite supplying tips needs to rapidly transmit information to a cueing satellite. “Any inefficiencies lead to poor use of the cued satellite resource, which is typically the most valuable,” Mason said.
As sensors gather more imagery, matching supply with demand becomes increasingly difficult. It’s a matter of understanding exactly what a sensor can see at a specific time and how quickly it can turn from one task to refocus on a new site.
Then, operators must consider financial issues particularly when satellite operators join forces to tip and cue.
“Have you already booked up your sensor? What happens if an order comes in later? Does it push out an order that’s already been committed to? How do you structure economic deals between the providers?” asked Peter Wilczynski, Maxar Intelligence chief product officer.
Spectral ‘fingerprinting’
Planet, which has about 200 satellites in low Earth orbit, gathers global imagery of Earth’s landmass with a resolution of about three meters per pixel almost every day with a product called PlanetScope. The company’s SkySats provide 50-centimeter views.
Meanwhile Tanager-1, Planet’s first hyperspectral satellite which launched in August, captures data from the visible to the shortwave infrared regions of the electromagnetic spectrum.
“The combination of these three layers allows for a global tipping system for detecting unexpected changes, high-resolution cueing within hours to zoom in on the finer details, and finally hyperspectral imaging for spectrally ‘fingerprinting’ what can be seen,” Mason said. “The tips can be customized from a generic pixel-change detector to more classical application-specific remote-sensing models. And the whole process can be automated by setting up a change-detection trigger that tasks a satellite to go capture the scene of that change.”
Nvidia edge processors on Planet satellites help analyze the imagery to generate tips that are “rapidly relayed to other satellites as cues, thereby drastically shortening the tip- to-cue time,” Mason said.
Zooming in and out
At Maxar, the company identifies change and provides frequent views of locations of interest by tipping and cueing high-resolution imagers on its Worldview and Worldview Legion satellites as well as partner spacecraft. In recent years, Maxar has forged agreements to task Umbra radars as well as Satellogic multispectral and hyperspectral sensors.
“We’re talking about operating multiple constellations as an integrated system,” Wilczynski said. “It’s particularly important because there are so many different types of sensors in orbit.”
Increasing the rate of collection and combining sensor observations allows Maxar to offer the National Reconnaissance Office, the National Geospatial-Intelligence Agency and other customers time-lapse sequences and imagery that zooms-in and zooms-out from events like fires or earthquakes.
Thanks to AI, low-resolution imagery of stationary objects can be paired with higher-resolution views. And multiple low-resolution images can be merged and enhanced to improve clarity through a process known as hyper-resolution. With high-resolution imagery becoming more prevalent, customers are clamoring for timely observations.
“For the next few years, our products are going to start being differentiated by currency more than they will be differentiated by resolution,” Wilczynski said. “People want up-to-date representations of the Earth, and they are willing to pay a premium for the best resolution and the best currency.”
Faster insights
At BlackSky, tipping and cueing is automated by the company’s AI-based software platform Spectra. When partners share tips, Spectra automatically cues the next satellite in BlackSky’s roughly 13-satellite constellation that is able to fulfill the request.
“This is all about getting faster insights into what’s happening,” O’Toole said.
For example, if HawkEye 360, a firm that maps radio-frequency emissions, wanted to confirm the location of a ship that had turned off its Automatic Identification System, it could share the coordinates with BlackSky through Spectra. BlackSky would then collect and share imagery.
Tipping and cueing isn’t limited to terrestrial observation, though. The speed at which requests can be processed in near-real time means companies like Australian startup HEO can share orbital tracks of space objects it wants BlackSky to observe.
Spectra can also be programmed to tip and cue based on the movement of vehicles or vessels.
“We use tips from third-party sources on the locations of barges and ships, for example,” O’Toole said. “Then, we automatically task satellites to collect images whenever these ships enter certain locations.”
Or if they are covered in netting.
This article first appeared in the May 2025 issue of SpaceNews Magazine with the title “Tip Off.”
