Someone sends you a photo on WhatsApp, you find one in an old folder, or you spot one in a listing, and you wonder: where was this taken? You open the file details, look for coordinates and find nothing. That is not the exception but the rule. Most photos you come across no longer carry usable GPS data.
That does not mean the location is lost. There are several ways to place a photo, from simple manual checks to AI. In this article we walk through five methods, in the order you should try them.
Why GPS data so often goes missing
A smartphone stores location in the photo’s EXIF metadata, as long as the camera app has permission. But as soon as a photo is shared, that metadata is often stripped. WhatsApp, Instagram, Facebook and Marktplaats remove location data on purpose, for privacy reasons. A screenshot of a photo never has GPS data. Photos from an old archive, a regular camera or an editing tool often lack coordinates too.
Sometimes the data is there but unreliable. Anyone who edits a photo can change the metadata, and a phone with a weak GPS fix may write a location from minutes earlier. So never trust EXIF blindly. To see what your own photos give away, read our article on the privacy risks of geotagging.
Method 1: check the EXIF data first
Always start with the simplest check. Open the photo in its original form, so not through a chat app, and look at the metadata. On a computer you can use the file properties, or our EXIF GPS viewer, which also shows the map right away. It takes ten seconds and sometimes you get your answer immediately.
If you see no coordinates, look at the rest anyway. The camera model, timestamp and time zone can help you test a story. A photo claimed to be taken “yesterday in Utrecht” with a timestamp from three years ago is already suspicious.
Method 2: read what is in the picture
If the metadata gives you nothing, look at the photo itself. Look for text: street signs, shop names, number plates, billboards, a bus line on a stop. Also look for Dutch features that give away the region: the colour of the bricks, the shape of the facades, the type of cycle path, the paving, the kind of road sign.
An example: you see a row of canal houses with stepped gables, a bridge with a narrow edge and a no-parking sign. That points to a historic city centre, and with the name of a café in the background you can narrow it down a lot. How this works and why it suits AI so well is covered in our piece on visual clues.
Method 3: search in reverse and compare with map imagery
Reverse image search (via Google Lens, Bing or TinEye) sometimes finds a photo’s source, for instance the original listing or an article where it appeared before. That mainly works for images already online, and far less for a private photo or a street corner without a striking building.
If you have a hunch about the area, compare the photo with street imagery in a map service. Look for fixed points: the corner of a building, the position of a lamp post, the shape of a junction. This is precise work, and it often takes half an hour per photo, but it is how investigators have worked for years. You can read more about this approach in our practical OSINT guide.
Method 4: use time, sun and shadows
Sun and shadows add information. The direction of shadows shows which way the camera was pointing, and together with the time you can estimate which facade faces south. The season helps too: leaves on trees, tulips in bloom or a frosty street rule out periods and narrow your search area.
This is a helper, not a solution on its own. You use it to confirm or reject a candidate location, not to find a place out of thin air. It works best as a complement to one of the other methods.
Method 5: let AI geolocation do the work
The fastest way is AI geolocation. A model like CosPlace is trained on millions of street images and compares your photo with an index of locations. It does not need to read text and does not need EXIF: it looks at the surroundings as a whole and returns a likely location within seconds.
For photos in the Netherlands this works especially well, because GeoPin is built on a nationwide street imagery index. The result is a probability, not a certainty, so here too: verify it with one of the earlier methods when a lot is at stake. Curious what happens under the hood? Read how GeoPin works.
Which method to choose when
A good routine is to always start with EXIF, and then pick the fastest method that gives enough certainty. A loose photo from a group chat can often be placed in seconds with AI. If it concerns a claim, a fraud report or a publication, combine two or three methods and keep your findings.
The nice thing is that the methods reinforce each other. An AI result that matches a street sign and a consistent shadow direction is far stronger than any of the three alone. That is how you go from “it could be” to “this is the place”.
Try it yourself
Do you have a photo without GPS data and want to know where it was taken? Upload it to GeoPin and see within seconds where the image comes from according to the model.