The Internet is full of diehard hobbyists, all quietly plugging away at their beloved pursuits on niche forums and personal blogs. Take astronomy, for example. On Reddit, there are more than 100,000 “stargazers” posting and discussing their best snaps of the sky. In the Flickr group for astrophotography, users have uploaded more than 86,000 photos of distant planets and stars.
This trove of images represents more than a few nerds’ online obsession, though. According to a new paper from researchers at three leading universities, those photos could be used — in theory, at least — to create an image of the night sky more accurate than the ones taken by the Hubble Telescope or the Sloan Digital Sky Survey.
In the words of the paper itself, harnessing the power of web hobbyists is “a holy grail” for astronomers.
How scientists plan to uncover that holy grail is, of course, a bit technical (as these things often are). But in an e-mail to The Post, the paper’s lead author, Carnegie Mellon’s Dustin Lang, basically laid it out this way: Large, professional telescopes are really powerful, but they still have to make some trade-off between area (how much sky they capture) and sensitivity (how detailed the capture is). We have very, very detailed images of tiny tiles of sky … and we have semi-detailed images of much bigger swaths.
Lang and his co-authors, however, have essentially developed an algorithm that can accurately combine hundreds of photos — even really amateur, point-and-shoot photos — into huge composite images that show large portions of the night sky in a high level of detail. To put it simply, Lang explains, “a bunch of small telescopes distributed around the world have capabilities that a single large telescope does not.”
Lang is quick to point out that his paper doesn’t contain any astronomical breakthroughs; he and his colleagues didn’t discover “anything that wasn’t already known about the sky.” But they did essentially discover, or at least advance, a fascinating new vein of research — one that could, with enough time and photos, change the way we see the night sky entirely. Here’s Lang:
In our paper we had to scrape the web for images, which means we didn’t get very many images, and we don’t necessarily even have the right to use the ones we found! But if we asked citizen-scientists to contribute to a project to observe a specific set of galaxy targets, we would probably get a whole lot more images, and we would be certain we were allowed to use their images … This work opens up the interesting space for amateur astronomers across the web to collaborate on focused observing campaigns.
To that end, Lang and his colleagues have launched a Web site that will collect and collate amateur pictures into an open-source sky map. In fact, if you go to astrometry.net, you can upload your own JPEGs or GIFs of the night sky, as well as explore photos submitted by other users. As of this writing, more than 3,200 people have joined the site; some have already posted tens of thousands of images. Even the Web’s most obscure hobbyists, it turns out, can be a force for mainstream good.