[David Schneider] was reading about recent discoveries of exoplanets. Simply put these are planets orbiting stars other than the sun. The rigs used by the research scientists include massive telescopes, but the fact that they’re using CCD sensors led [David] to wonder if a version of this could be done on the cheap in the backyard. The answer is yes. By capturing and processing data from a barn door tracker he was able to verify a known exoplanet.
Barn Door trackers are devices used to move a camera to compensate for the turning of the earth. This is necessary when taking images throughout the night, as the stars will not remain “stationary” to the camera’s frame without it. The good news is that they’re simple to build, we’ve seen a few over the years.
Other than having to wait until his part of the earth was pointed in the correct direction (on a clear night) at the same time as an exoplanet transit, [David] was ready to harvest all the data he needed. This part gets interesting really quickly. The camera needed to catch the planet passing in between the earth and the star it revolves around (called a transit). The data to prove this happened is really subtle. To uncover it [David] needed to control the data set for atmospheric changes by referencing several other stars. From there he focused on the data for the transit target and compared points across the entire set of captured images. The result is a dip in brightness that matches the specifications of the original discovery.
[David] explains the entire process in the clip after the break.
28 thoughts on “Astrophotography And Data-Analysis Sense Exoplanets”
I am astonished by this. I think I know my next project now once I’ve finished writing my novel.
I don’t know what’s more surprising to me: That someone tried this, or that it actually worked. Awesome project, David. Hack of the Day I’d say.
I am duly impressed.
Would the next logical step be doing this the crowd sourcing way, in essence giving out small patches of sky to observe?
Could the software used be made self explanatory and user friendly enough for that to work?
That is so cool!
As I am working on an rotator for antennas or cameras with closed loop position and automatic orientation finding with a magnetometer by my self at the moment I will try this out for sure the next time I am not at the light polluted city.
I also like the idea of making this in a way of crowd sourcing. Basically SATNOGS (same hardware with antennas replaced for cameras) for astrophotography.
Using many cameras pointed at the same object may allow even greater accuracy.
spectrum.ieee.org covered this before HaD. Not complaining, before people leap to Mike’s defense, but just letting people know because if you enjoy HaD, you will most certainly enjoy spectrum (but more science-y than hack-y).
Just another good site for your bookmarks/rss. :)
Yep, David is an editor at Spectrum and we’re featuring his work and giving him the credit. I’m not sure what your point is.
I think his point is abundantly clear if you read the actual words and take them at face value. You sound defensive.
He takes a uses a lot of words to say “this is covered at this other site if you want more information.” All the dancing around the point and “not trying to accuse” nonsense is unnecessary, and that he said it despite this fact makes it sound like he IS trying to imply something. It reads like a more polite/sane version of the old conspiracy theorist trick of bringing up some crazy idea and saying “not saying I believe it, just asking questions.” Whether that’s intentional or not, it’s how it came across.
Absolutely. I thought the HaD comments were better than this…apparently not. Even the author: “I’m not sure what your point is”, what the hell? My point is that MANY people who come here would really enjoy reading Spectrum. Come here and read awesome articles. Go there and read awesome articles.
People are too quick to blithely ignore the literal meaning of what I wrote and just as quick to join in on the increasingly circle-jerky HaD comments…comes with the new font, I guess.
Pls editable comments, etc, etc. It’s almost 2015.
> Pls editable comments, etc, etc. It’s almost 2015.
You want a required login system? That’s how you get a required login system.
Since I can’t reply to Brian, because apparently the comments have gotten too deep or something I dunno, I’ll reply to myself.
You absolutely can do editable comments without a login system, by storing a token for the comment server-side, and then also storing that token with HTML5’s local storage. There are plenty of js libraries that combine cookies + flash data storage + local storage to provide compatibility with a very wide range of browsers. There’s also 2ch style image-board deletion (and on some but not all, editing) where there is a ‘password’ field when you make a post/comment and if you want to go back and modify your post/comment, you just have to remember the string that you used as your ‘password’ for that post.
I love HaD, and I’m fine with the comments as they are and always have been. My asking for it was mainly intended as a joke, because every second comment seems to be asking for them.
Now that we’re somewhat done with the major changes to the layout (see the improvement over the last two weeks? Isn’t it great now?), we can start tackle the long tail.
The point is that people like me who were not aware of that website can now go check it out. And if you miss the point of a comment then you are a moron because he literally summarizes his three sentence comment with the third sentence. If you tone down your defensiveness you’ll be able to interpret more of the world around you.
Thanks to all for your kind remarks on this project. I’m still pretty astonished myself that it worked.
Iageos: Forgive me for raising concerns before you’ve even tried, but I doubt very much that the kind of tracking mechanism used to rotate an antenna would be much use for something like this. Although a barn-door tracker is a pretty simple piece of hardware, it can provide remarkably good tracking if built properly. For example, I was easily able to see periodic variations in the threading of the threaded rod used to drive apart the “doors” (and was able to reduce those variations by swapping out the hardware-store rod I first used for a better-made one). I can’t imagine that a typical antenna tracker could provide sufficient accuracy in tracking stars through a telephoto lens. You don’t need super-high tracking accuracy for this project, but if too rough it’ll make the data processing really difficult.
Thanks for the input!
It is planned as a verry sturdy construction with DC motors with an gear ration of 1:173 and a hall sensor readout of 48 counts per revolution giving a theoretical resolution of 0.0433 deg. Because of direct mount the motor movement should be correspondent to the movement of the camera although alignment is still an issue even with lacercut parts). So yes, an average antenna tracker will have problems.
The setup is compareable with a dobson telescope (because I want to be multipupose), but with ball bearings instead of sliding ones (to reduce slip stick) i.e. altitude and azimuth so I can use pyephem ( http://rhodesmill.org/pyephem/ ) for a kind of automatic tracking with PI(D).
Please correct me if I am wrong: Your field of view is somewhere arround 3 to 5 degrees, correct (to have a ballparknumber for accuracy)?
What software did you use for analysis of the data?
Exactly right: the field of view with the 300-mm lens I used is just about 3 degrees x 5 degrees.
I used Iris to do the analysis. You can find a link to it and some other very helpful online resources in my IEEE Spectrum article.
Can’t you just align the images in software by using some reference stars? I mean how does the tracking variation influence the brightness data? And If it has an influence you can just subtract the variation seen in the aforementioned reference stars to get a correction profile I would think. Although someone would have to test to be sure that works.
Actually, do you need to track at all? Maybe you could evaluate the stripes created by a long exposure and that will be detailed enough to determine the brightness variation with sufficient precision. Assuming they don’t run over each other for the targeted star of course.
Doing this without tracking would be a feat. Consider this: The Earth rotates 15 degrees per hour. The transit I measured takes almost two hours. When I did this, I took images for almost 3 hours. During that time, the star of interest moved almost 45 degrees in the sky. The field of view with a 300-mm telephoto is only a few degrees, so you need tracking just to keep the star in view for the requisite amount of time.
I suppose you could try this with a much wider lens. But that would probably introduce a host of complications (including the one you mentioned). So yes, in principle, you could, I suppose, do this without tracking with a sufficiently wide-angle lens. But I’m not optimistic.
I thank you for your detailed reply, as you must have gathered I did not do the calculations so it’s nice to see the data and why this would be rather hard to do. And if your explanation wasn’t enough I can add another problem, the tracks would not be straight lines over longer time so to even get that brightness info on a single star from an image might be a bit tricky too. So overall it would indeed be silly to make life so difficult for yourself with no apparent benefits from doing it the hard way.
I’ve been wondering if this kind of thing would be possible on a back yard budget.
Recently got into astrophotography myself, capturing 4 of Jupiters moons was my best moment so far
Have you considered using a DSLR with a telescope on a motorised mount to capture more light?
WIll be following your work with great interest.
Yup, definitely a back-yard budget project here. If you don’t count the cost of the DSLR camera (which I got used on the cheap anyway) or the computer (who doesn’t have one?), it should cost less than $200 for the various bits and pieces.
Using a telescope and motorized mount (particularly a guided mount) I’m sure would be terrific for astrophotography. But you don’t really need more light to detect an exoplanet. Indeed, I had to turn the ISO way down to avoid saturating my camera’s sensor with the long time exposure I used (50 seconds). The long exposure was to average over scintillation (twinkling). All to say, for this particular exercise, you don’t need a telescope, a motorized equatorial mount, or even particularly clear skies.
Yes, yes, yes! This is exactly the kind of thing that needs to be crowdsourced, can you imagine how many exoplanets could be found by a million backyard hackers! And the best part is if you find a new planet you get to name it! Planet Kickass here we come!
Planet Hackaday. Now, that would be awesomeness, redefined.
That is a brilliant idea, scale it up and have a sky watcher cluster… thousands of DSLR’s looking upwards, tracking the sky, working together to find planets, track asteroids, satellites, etc….
Instead of using a long exposure, is it possible to just take many photos in succession (e.g., a video) and average across each frame?
I’m no astrophotography expert, but something like that might work. (Search on “astrophotography” and “webcam,” and you’ll find lots of people using that general approach.) That said, you’d likely still need a star tracker for the reasons I described in replying to Whatnot above.
Please be kind and respectful to help make the comments section excellent. (Comment Policy)