New software shows promise for coping with photobombing satellites

An Earth-orbiting satellite leaves a streak as it crosses in front of the Hubble Space Telescope and its Advanced Camera for Surveys during an observation of “The Mice,” a pair of interacting galaxies catalogued as NGC 4676. With satellite mega constellations rapidly growing in low-Earth orbit, researchers have developed software that makes it easier to detect, and then erase, such photobombing satellites from Hubble observations. Image: NASA, ESA, STSci

Mega constellations of Earth-orbiting satellites, like the Starlinks now providing space-based internet service, are growing day by day, raising concern about interference with astronomical observations as the spacecraft increasingly photobomb observations, leaving white streaks across otherwise pristine images.

Last year, 10 percent of the images captured by the Hubble Space Telescope’s Advanced Camera for Surveys, the instrument with the widest field of view, were affected.

But researchers at the Space Telescope Science Institute in Baltimore, Maryland, have developed software based on an image analysis technique known as the Radon Transform to identify satellite streaks in ACS observations so they can be digitally removed.

“We developed a new tool to identify satellite trails that is an improvement over the previous satellite software because it is much more sensitive,” said Dave Stark of STScI. “So we think it will be better for identifying and removing satellite trails in Hubble images.”

Hubble observations typically require multiple exposures of the same target. As a satellite streaks through the field of view, it can leave a trail across one exposure but not in the next. Stark and his colleagues developed software that identifies affected pixels and flags them so they can be “subtracted” from the combined images.

“When we flag them, we should be able to recover the full field of view without a problem, after combining the data from all exposures,” Stark said..

NASA says the software is “ideal for identifying and characterising linear features in an image because it sums up all the light along every possible straight path across an image. This approach combines all the light from a satellite trail, making them ‘pop out’ in the transformed image, even many of those that are very faint in the original image.”

The new software is 10 times more sensitive than earlier techniques, identifying roughly twice as many streaks.

“We have a toolbox of things that people use to clean Hubble data and calibrate it,” Stark said in a release. “And our new application is another tool that will help us make the best out of every Hubble exposure.”