A research study. One photo of the Moon among stars, and we work out where you stood.
Satellite positioning tells your phone where it is, but it cannot prove it to anyone else. Authentication is being added to the signals themselves — Galileo now signs its navigation messages — but that is not the weak point. A receiver works out where it is from the differences in arrival time between satellites, and those differences can be forged by relaying genuine, properly signed signals with the delays you choose. The message is authentic; the geometry it implies is not. We are testing whether the sky can do the job instead.
The Moon is close enough, about 384,000 km, that it appears in a slightly different place against the background stars depending on where on Earth you stand. The shift reaches about 1°. The stars fix your camera's orientation; the Moon's offset against them fixes your position. One photograph, one position — no satellites and no network.
None of the astronomy here is new. Navigators worked out where they were from the Moon and the stars long before satellites, with a sextant and an almanac. What is new is the camera: an ordinary phone sensor resolves the sky finely enough to do it automatically, in seconds, with no instrument and no training.
And what is new is doing it without giving yourself away. As things stand, a photograph can only prove where you were by revealing where you were, which is a poor bargain — you hand over your coordinates, your sky and your timestamp to whoever is asking. Zero-knowledge cryptography breaks that link. The arithmetic that turns a photograph into a position is ordinary arithmetic, and it can be carried out inside a proof: what you hand over is then a single statement — the person who took this photograph was inside this region at this time — while the image, the stars and your exact coordinates never leave your phone. The verifier learns that the claim is true and learns nothing else.
That is where this is going, and it only works if the measurement underneath is good enough to be worth proving. A proof cannot make a position more accurate than the photograph it came from. So the measurement comes first: how accurate is this in practice, on ordinary phones, from many places on Earth?
The accuracy depends on how high the Moon sits in your sky, which depends on your latitude. Our own photographs come from three places, so the result currently rests on a handful of frames from a narrow band of the planet. A photograph from your latitude is worth more to this study than another from ours.
Honest expectation: this is hard. In our own sessions — shot by the person who wrote the instructions — most frames yield nothing, usually because they caught too few stars. We have built an instant check so you find that out at the tripod rather than afterwards, but please do not be discouraged if your first attempt does not work. A failed frame is still useful to us and still counts as taking part.
Already have one? If you happen to own a photograph of the Moon with stars around it, please send that too — you do not have to go outside tonight. The one requirement is that the file still carries its exact capture time; without that the photograph cannot be used at all.
Short version: your photograph, when it was taken, and roughly where you were. That is the whole dataset. We do not ask for your name and you do not make an account. The privacy page lists every field we store, how long we keep it, and how to have it removed.
That is a sensible instinct, and this study is literally about not having to. Turn geotagging off in your camera before you shoot — then the file carries no coordinates at all — and just type your city when you upload. We will look up its centre and use that.
It costs us almost nothing. Our method currently lands within about 100 km, so a town centre is far more precise than the thing we are measuring. Being accurate to a few hundred metres does help us pin down how good the method really is, so it is welcome if you are comfortable with it — but a town name is a real contribution, and we would much rather have your photograph on your terms than not at all.
You can leave us an email address, and when the write-up is ready we will send you a copy of the paper — so you can see what your photograph went into. It is optional — skip it and your contribution still counts, and nothing else changes. We use the address for that and nothing else: no list, no forwarding, no publishing.
Everything else — the frames kept from the instant check, the session cookie, the contributor id, the weather lookup, and how to have your contribution deleted — is on the privacy page.
Please photograph from your own balcony, garden, or somewhere you would normally be after dark. Do not travel to an unfamiliar place, and do not go anywhere alone at night for this. A suburban sky with some light pollution is perfectly usable — several of our own successful frames were taken from a balcony in a city.