Skip to Content

How Crowdsourced Astrophotographs on the Web Are Revolutionizing Astronomy

Astronomers have long known that combining the data from several astrophotographs can reveal dramatically more detail about astrophysical objects. So what will they discover by combining all the astrophotographs on the Web?

Astrophotography is currently undergoing a revolution thanks to the increased availability of high quality digital cameras and the software available to process the pictures after they have been taken.

Since photographs of the night sky are almost always better with long exposures that capture more light, this processing usually involves combining several images of the same part of the sky to produce one with a much longer effective exposure.

That’s all straightforward if you’ve taken the pictures yourself with the same gear under the same circumstances. But astronomers want to do better.

“The astrophotography group on Flickr alone has over 68,000 images,” say Dustin Lang at Carnegie Mellon University in Pittsburgh and a couple of pals. These and other images represent a vast source of untapped data for astronomers.

The problem is that it’s hard to combine images accurately when little is known about how they were taken. Astronomers take great care to use imaging equipment in which the pixels produce a signal that is proportional to the number of photons that hit.

But the same cannot be said of the digital cameras widely used by amateurs. All kinds of processes can end up influencing the final image.

So any algorithm that combines them has to cope with these variations. “We want to do this without having to infer the (possibly highly nonlinear) processing that has been applied to each individual image, each of which has been wrecked in its own loving way by its creator,” say Lang and co.

Now, these guys say they’ve cracked it. They’ve developed a system that automatically combines images from the same part of the sky to increase the effective exposure time of the resulting picture. And they say the combined images can rival those from much professional telescopes.

They’ve tested this approach by downloading images of two well-known astrophysical objects: the NGC 5907 Galaxy and the colliding pair of galaxies—Messier 51a and 51b.

For NGC 5907, they ended up with 4,000 images from Flickr, 1,000 from Bing and 100 from Google. They used an online system called that automatically aligns and registers images of the night sky and then combined the images using their new algorithm, which they call Enhance.

The results are impressive. They say that the combined images of NGC5907 (bottom three images) show some of the same faint features that revealed a single image taken over 11 hours of exposure using a 50 cm telescope (the top left image). All the images reveal the same kind of fine detail such as a faint stellar stream around the galaxy.

The combined image for the M51 galaxies is just as impressive, taking only 40 minutes to produce on a single processor. It reveals extended structures around both galaxies, which astronomers know to be debris from their gravitational interaction as they collide.

Lang and co say these faint features are hugely important because they allow astronomers to measure the age, mass ratios, and orbital configurations of the galaxies involved. Interestingly, many of these faint features are not visible in any of the input images taken from the Web. They emerge only once images have been combined.

One potential problem with algorithms like this is that they need to perform well as the number of images they combine increases. It’s no good if they grind to a halt as soon as a substantial amount of data becomes available.

On this score, Lang and co say astronomers can rest easy. The performance of their new Enhance algorithm scales linearly with the number of images it has to combine. That means it should perform well on large datasets.

The bottom line is that this kind of crowd-sourced astronomy has the potential to make a big impact, given that the resulting images rival those from large telescopes.

And it could also be used for historical images, say Lang and co. The Harvard Plate Archives, for example, contain half a million images dating back to the 1880s. These were all taken using different emulsions, with different exposures and developed using different processes. So the plates all have different responses to light, making them hard to compare.

That’s exactly the problem that Lang and co have solved for digital images on the Web. So it’s not hard to imagine how they could easily combine the data from the Harvard archives as well.

One final point is that this project is open to anyone who takes astrophotographs. You can submit them to, where they will be combined into an open source sky map. In return, users receive annotated versions of their images as well as the combined picture of the sky that it relates to.

It wasn’t so long ago that the age of the amateur astronomer looked to be coming to an end because a new generation of huge telescopes and space-based observatories were producing data that amateurs could never hope to get.

All that is changing now. Long live the amateur astronomer!

Ref: Towards building a Crowd-Sourced Sky Map

Keep Reading

Most Popular

Russian servicemen take part in a military drills
Russian servicemen take part in a military drills

How a Russian cyberwar in Ukraine could ripple out globally

Soldiers and tanks may care about national borders. Cyber doesn't.

Death and Jeff Bezos
Death and Jeff Bezos

Meet Altos Labs, Silicon Valley’s latest wild bet on living forever

Funders of a deep-pocketed new "rejuvenation" startup are said to include Jeff Bezos and Yuri Milner.

conceptual illustration showing various women's faces being scanned
conceptual illustration showing various women's faces being scanned

A horrifying new AI app swaps women into porn videos with a click

Deepfake researchers have long feared the day this would arrive.

ai learning to multitask concept
ai learning to multitask concept

Meta’s new learning algorithm can teach AI to multi-task

The single technique for teaching neural networks multiple skills is a step towards general-purpose AI.

Stay connected

Illustration by Rose WongIllustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at with a list of newsletters you’d like to receive.