Slashdot Mirror
Who Needs NASA? Exoplanet Detected Using a DSLR
Iddo Genuth writes Until 20 years ago even the best telescopes in the world could not detect a planet outside our solar system. Now, with the aid of a basic DSLR, low cost lens and some DIY magic, you just might be able to "see" ET's home planet for yourself. Your DSLR can do much more than just take a few nice portraits or the occasional vacation photos – if you have some DIY experience (O.K. a bit more than just "some"), you might be able to repeat what David Schneider was recently been able to do — that is, building his own planet finder using only inexpensive photo gear, low cost electronics, the right kind of software and a lot of patience. Although Schneider was "only" able to rediscover an already known exsoplanet (some 63 light-years away from us), what he did — and more importantly how he did it — might allow planet hunting to become closer to SETI@home than NASA's 550,000 million dollar Kepler space telescope project.
"550,000 million dollar Kepler space telescope"
I think you're about 3 orders of magnitude too high.
The actual article is here, on ieee.og
http://spectrum.ieee.org/geek-life/hands-on/diy-exoplanet-detector?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+IeeeSpectrumFullText+%28IEEE+Spectrum+Full+Text%29
The more complete source article
Prediction for end of Universe #42: Fencepost error in Quantum_bogosort.cpp
Well, check out NASA's website. Their headline is 'Who Needs Low-Budget DSLR-Wielding Amateurs?'
This is a really cool project, with a terrible headline. Without NASA (or perhaps the ESA, or whatever space organization first found this exoplanet), David Schneider wouldn't have been able to look up the timing for the planet's transit. He wouldn't even have know to try taking pictures of that particular star. He'd have to take a lot more photos over a much longer time, over a much bigger area of the sky, and run a lot more image comparision software for a lot longer, before he's have found that transit.
a smartphone and an app!
Well ok so some Italian already discovered Jupiter's moons, but if they hadn't, you could!
To actually *go* there!
Want a shallow depth of field? Think you need a large lens / fast lens? Why? When the computer can take multiple photos at different focal lengths and calculate the depth properly. Why twiddle with zoom and f-stop?
Because by the time my camera gets around to taking several photos at various focal lengths, the fast moving subject I was trying to take a picture of already left the scene. But I have the opposite problem (since I can fake a shallow depth of field in photoshop) -- I have to open the lens all the way on my P&S in any kind of low-light situation since the sensor is too noisy at high ISO's, so I end up with a shallower depth of field than I wanted.
Want a fast lens? Point and shoots go down to f1.8 now. I can photograph the craters on the moon with a pocket sized G7X without a tripod FFS. How often do you try to detect planets in other solar systems?
I want a fast lens *and* a larger, lower noise sensor than I can find in most P&S cameras. But I don't use that fast lens for taking pictures of planets or the moon.
Want to shoot wildlife at a distance? Think your DSLR is the best option? Think again, you should be using a higher pixel camera at a lower zoom, because you'll have difficulty tracking the moving object at high zoom with that lumping great lens.
Right, all those sports photographers that use the big $8000+ 300mm telephoto lenses could save some money and just get a point a shoot with more pixels than his 20MP DSLR and he can just crop down the pictures to give him a nice 3MP shot of the winning touchdown. Pixels are pixels, right? The 1/2" sensor on a 20MP point-and-shoot is just as good as the 35mm sensor on his DSLR, right? And what possible difference could there be between an $8000 lens and the lens on a $500 P&S?
Want to shoot movies? Do you see any pros using DSLRs? They use a Red or similar, not a Canon EOS.
Here's a list of 30 movies and TV shows that have been shot in whole or part on Canon or Nikon DSLRs:
http://www.imdb.com/list/ls059...
"Honey, I just discovered a planet and I'm going to name it after you. An entire planet."
Table-ized A.I.
If the article is correct it's only useful for checking transit events. The article talks about "a target star". "A star" has a low probability of having a planet in an orbit that gives transit events. Depending on the orbit of the planet it is next to 0 to 10%.
That is why the expensive systems check many stars at once. Kepler monitors 145,000 stars at once
That doesn't mean much. NASA always lacks money to do awesome stuff. So I feel that cost is a good metric here:
Kepler monitors 145,000 stars and costed $600,000,000. That means $4,138 per star
This costs $300 + $100 + some work = +/- $500 per star. If the same software could be used then it might be possible to check more stars with a single DSLR. Maybe even a hundred.
That means $6 per star (naively assuming his setup is sufficient for large scale deployment)
I say: if and when Kepler dies we should stick a couple of them on every observatory we have and have full time monitoring across the planet. False positives drop with more data.
False negatives would still require different detection methods since most exoplanets don't transit. However if we have this set up we might be able to add thousands of spectrometers to it to detect Doppler shift events. Then the only planets with low detection chance are the ones where we look directly on the top (or bottom) of their orbit.
Well, I might have a way, but it only works on a semi spherical planet in a vacuum.
I really want to think that this was a great feat of hardware hacking, but the title of the summary is terrible. The capabilities of space telescopes are designed to overcome the obstacles which plague our earth-bound ones.
Saying that this is a viable replacement for the data coming from a source with lesser disturbances is just undermining the work of a lot of people.
I must admit in skeptical. It is very easy to "see" something that you know is there in data, eliminating runs that you "know" are wrong, etc. Until this is reproduced by someone with similar equipment I will put this down to a fluke
One cannot escape the fact that bigger apperture telescopes can record fainter
stars, and/or perfom the photometry of bright stars with more precision than a simple camera.
To detect exoplanets one needs both large samples of stars recorded as continuously as
possible over several years and high precision photometry. Besides being cheap, the advantage
of a small camera is than the field is larger. But with a larger telescope in space like Kepler one
can target regions of the sky with density of stars optimal for the CCD/camera combination, and
observe continuously for months with the same instruments, which is crucial for differential
photometry. Thousands of amateurs worldwide detecting as many new exoplantes as Kepler
would face the problem of coordinating the analysis of huge amounts of heterogeneous and
incomplete data (due to day/night and weather interruptions in differently dark and transparent skies).
The real question is wether crowdsourcing planet detection is cheaper for global economy at equal scientific return than with state sponsored research. Perhaps the most important benefit of such an
activity is educational and promotional for research in general.
Wow... You are wrong on almost every count in this post. Impressive.
Seriously, a DSLR sensor is 35mm because the film it replaced is 35mm.
Mostly wrong. Only high-end professional-oriented DSLRs have 35mm-sized sensors, the majority use smaller APS-C sized sensors or something in that ballpark.
"Slashdot - News and Chat Sites Deviant". (Click "homepage" link above for details).
One reply has already gutted most of your points, so I'll work on the ones he didn't address:
Seriously, a DSLR sensor is 35mm because the film it replaced is 35mm. It's not an engineering choice, it was done so that owners of expensive lenses could use them on the new digital cameras.
Except that not all DSLRs are 35mm, in fact most of them weren't, and most of them still aren't. They were APS-C sized. Or in the case of 4/3rds they were even smaller and not even the same aspect ratio. Yet all the lenses magically still worked.
35mm has enough engineering decisions behind it,
Its not sized large enough to take in a normal range of light, and yet small enough to keep the lens size down so big zoom lenses are possible. It's sized for historic reasons, making it too big and thus limited.
And yet there's no definition for normal range of light and it seems every year there are changes to sensitivity. Are you saying that it was initially not large enough? What about now when cameras go to ISO25600 and beyond?
Lots of things are better done by digital calculation than lenses.
False. Leaving aside the technicalities of deconvolving and correcting for aberrations at different zooms and focus point, there's a much more fundamental saying: "Shit in. Shit out." Every correction make negatively affects an image that could have been right to begin with given how well lenses are understood.
How often do you try to detect planets in other solar systems?
Are you asking how often a man does his hobby? Who are you to question?
Want multiple lenses for those oddball occasions, go with Micro 4/3 like a lot of pros are doing now.
Why, and Who? Seriously I love my EM1 as much as the next person, but you'd be hard pressed to actually point out a pro using micro-4/3rds for anything other than an Olympus promotional video. As for oddball lenses, I think you'll find both the EF mount and the F mount will have a far wider and more oddball selection than anything 4/3ds.
As sensors have gotten better, and lower noise, the sensor in the DSLR has not shrunk in size because of the lens.
No. Quite the opposite. The sensor size has increased. The coupled larger format with the newer sensor technology now makes it possible to take photos that were previously impossible without fancy lighting and careful setups. Upping the sensitivity is no excuse for making things smaller when the benefits can be had across the board, especially since small sensors struggle to get depth of field down.
Big zoom lens are unworkable in DSLR because of the size of the sensors, they would simply be too long.
And yet they have been built and are in active use.
Likewise the size of the sensor is a big problem itself, its slow electrically.
And yet they are used in fast paced sporting situations and 60fps HD video.
So you end up with a camera with worse performance in normal situations, and designed for non-real world situations.
So a camera designed not to take pictures? I'm not sure I follow here.
Trying to find planets literally is what these camera are useful for!
No trying to find planets are what carefully calibrated and cooled CCDs are useful for. Before someone can use a DSLR for anything as technical as this they need to move heaven and earth to properly quantify how their system reacts as sensors used for photography are inherently non-linear trading off accuracy for other aspects. I know someone who uses his Canon DSLR for gauging the variability of stars. He literally recorded months of data before he had something useful enough that he could calibrate his camera for use on a single target.
The fact this works at all is amazing and a testament to the people involved, and definitely nothing to do with technology.
The gp displays the standard ignorance of a novice photographer, who doesn't even know that he knows nothing about photography because his hardware allows him to take pictures he deems "good". There's a reason there are professional photographers out there that get paid well and use "pro" equipment, but until gp tries to take pictures outside the limited scope of a P&S, he won't realize how limited he is. Should he use a pro-sumer or better grade camera, he may be disappointed in his pictures not being "better" than the P&S, that's not a fault of the equipment but more his failure to understand how to use that equipment to get great pictures. And even then, great pictures don't just happen automatically, it is luck or many tries, and sometimes significant post-processing before a great picture is realized. To be fair, sometimes that is also achievable with a P&S.
The cesspool just got a check and balance.