• OSINTSUM
  • Posts
  • How to Geolocate Imagery using OSINT

How to Geolocate Imagery using OSINT

Using clues held within an image file to identify where it was captured

Step 1: Check for EXIF data

An Exchangeable Image File Format (EXIF) is a data set which provides metadata about media files such as images, videos, or audio.

This includes details such as the type of camera it was shot with, the shutter speed, and lens type. Even more helpfully, it may also include data on the time and date the image was taken, and the GPS coordinates of where it was taken.

An example of EXIF data from a media file. Image Credit: thomocochrane

Caveat: Don’t get too excited, while this is the god-tier indicator for geolocation, often it will not exist.

Social media platforms routinely wipe EXIF data from imagery to protect the privacy of users, and even if the file isn’t from social media there are a host of other reasons it may not exist: the camera wasn’t set to record it or its GPS was disabled when the image was taken, the user wiped it, the file is corrupted, or the file may be a copy of a copy of a copy.

To check whether EXIF data exists, simply right click on the file and select “Properties” on a PC or “Get Info” on a Mac. If your computer doesn’t have this feature, you can go to a site like www.metadata2go.com which allows you to upload files and checks them for EXIF data for you.

If you don’t want to download an image you’ve found online, you can try copying and pasting the URL of the image into a site such as www.exif-viewer.com, and this will check for the existence of EXIF data for you.

If there is no EXIF data stored on the image file, the next step is to conduct a reverse image search and see if any other copies of the image, or similar images, exist online that can give you more clues as to where it was taken.

Sites that allow you to do this include:

Results from a reverse image search I conducted on TinEye

Step 3: Look for visual indicators

If your reverse image search is fruitless, try and identify potential clues within the image itself. These include objects, environment, and text.

Internal images

If the image has been taken inside (i.e. the external environment is not visible), it is unlikely you will be able to narrow it down much further than a specific country unless you can identify the actual room/building the image was taken in. This is an important point to make to any customers asking you to geolocate an internal image as part of managing their expectations.

Features to look for in internal images include:

Plug sockets

There are approximately 15 different models of plug socket in use across the world, with the majority of countries only using one or two of these models. Therefore, if the image shows a plug socket in a high enough resolution to identify which model of plug socket it is, you can narrow your search down to a small collection of countries. To see which countries use which kind of plug socket, look at our guide to geolocating images using plug sockets.

Objects

If you can see any objects in the image, such as appliances, branded product packaging, or toys, see if you can identify their make and model. Once you know the make and model of the object, you can look up which countries it was sold in, narrowing down your search. You will also be able to find out when it first went on sale - informing you of the earliest moment in time the image could have been captured.

One way of doing this is to do a reverse image search on the part of the image that shows the appliance. The majority of reverse image search engines now include a built-in cropping tool, but if your preferred one doesn’t just manually crop the image and upload that instead of the full image.

The initial reverse image search on the entire image brings up other copies of the image present online

Narrowing down the search to an object present in the image brings up examples of identical or similar objects from different images

Another way to do it is to speak to an expert. For instance, if I was trying to identify the make and model of an oven in an image, I would consider speaking to a technician at an appliance repair shop to see if they recognise it.

You can also take advantage of forums - you would be surprised at how niche some people’s hobbies and interests are. Find a forum, Facebook group, Discord, or other online community dedicated to the type of object or appliance you have identified and upload the image there to see if anyone is able to identify it for you.

Case study: Don’t Fuck with Cats

Amateur investigators attempting to locate a criminal who had been posting videos of animal abuse online spotted a yellow vacuum cleaner in one of his videos.

They asked about the vacuum cleaner on a forum dedicated to vacuum cleaners (I told you people have some niche hobbies!), and the forum members quickly identified it as a model only sold in North America - significantly narrowing down their search radius.

This incident was covered in the Netflix documentary Don’t Fuck with Cats, which I highly recommend if you are into OSINT (provided you don’t have a weak stomach!) as it has a number of great examples of OSINT analysis in it.

Note: Remember it is possible to transport objects overseas, so seeing an object that originated in a certain country in an image does not mean that image was definitely taken in that country. But, it is a good start and can be used to corroborate other evidence you find in the image.

Text

Look for any text you can find in the image - on packaging, products, letters/envelopes, credit cards, clocks, clothes, calendars etc.

If the text is in a foreign language or obscure alphabet, use Google Lens to identify what language/alphabet it is written in, and then translate it into your language.

Run any proper nouns in the text - names of schools, companies, products etc - through a search engine to find out more about them.

External images

If the image was taken outside, it may be possible to narrow down its location to a five-figure grid reference, the direction the image was taken from, and even the time and date it was captured.

Features to look for in external images include:

Landscape

Mountains, coastal features, plains, skyline etc are all useful for geolocating imagery. Straight away they narrow down the potential locations it can be taken without having to do any complex analysis (i.e., if there is a coast line the image was obviously taken at the coast and if there are skyscrapers it was obviously taken in a major city.)

Flora and fauna

Animals and plants tend to only inhabit certain areas of the globe, meaning that if they are visible in the image there is a good chance it was taken somewhere the animal or plant is native to. For instance:

  • Palm trees grow in tropical and sub-tropical regions, so an image containing palm trees is likely to have been taken in one of those regions.

  • The further north you go in the Northern Hemisphere, the greater the proliferation of evergeen/coniferous trees. The further south you go, the more deciduous trees.

  • Polar bears only live in the wild in the Arctic, so if you see an image with a polar bear in the wild in, it was almost certainly taken there.

Below is a map showing the general land biomes across the world

Climate

Clues about the climate in the image will help locate where it was taken. Look for weather clues such as sun, snow, wind, rain etc, as well as the vegetation, style of roofs, and the clothing people are wearing.

You will be able to narrow down the potential location of the image using climate evidence much further if you can also work out the time of year the image was taken.

Buildings

Architecture is often region-specific, meaning even if you can’t identify the exact building you are looking at, you may be able to tell where it is by the style of the building. As is often the case, Google Lens is your friend here.

Cars/Roads

Cars give away a goldmine of geographic information.

Look at which side of the road the cars are driving on. If they are driving on the left, it will narrow your search down to 75 countries, and 165 if they are driving on the right. For a full list of which countries drive on each side, click here.

If you can make them out, registration/number plates and country stickers on cars will tell you the country (and in some places the region) the car comes from.

Bear in mind that cars have the ability to travel, and therefore any clues about the origin of cars in images will only act as evidence of the image’s location and not definitive proof.

Check the road and its surroundings for markings, cats eyes, lamp posts, signs, bollards, traffic lights, speed cameras etc. All of these things will give a clue to where the road is located, and, as roads are fairly immobile, identifying the location of the road will give you fairly solid proof of where the image was taken.

Humans

In the West, populations are becoming increasingly diverse, and therefore it can be tempting to write off the ethnicity of subjects in an image as a reliable indicator of location. However, some areas are still mono-cultured, and therefore if the majority of people in an image are of a certain ethnicity, this can still be a useful identifier.

Look at the clothes the people in the image are wearing. As a general rule the warmer the clothes they have on the further from the equator the image was taken, but you will need to factor the time of year into this equation as well.

Contrails

If you can see contrails in the sky, see how high up they appear. The lower the contrails, the closer to an airport the image is likely to have been taken.

Colonialism

If you can see road signs in a certain language, but all the other evidence in the image points to it not being taken in the country that language originated, it is possible it was taken in a country that was once invaded by that country.

For example, many countries in the Bulge of Africa speak French.

If the camera is facing in a northerly direction, and the sun is visible in the image, it was taken in the Southern Hemisphere, and visa versa - instantly eliminating half of the world from your search radius.

If the image is looking in a northerly direction out to sea, then the image will have been taken on a southern coast of whatever country it was taken in. You can apply this logic to whatever direction your image was taken from.

Step4: Work out which direction the image was taken from

EXIF data

The EXIF data on the file may tell you which direction the camera was pointing when the image was taken. If it is there, it will be under Exif.GPSInfo.GPSImgDirection. It is highly unlikely this information will be present if the GPS location itself isn’t there, but always worth a check.

Shadows

Long natural shadows (shadows cast by the sun rather than artificial lighting) indicate the image was taken early in the morning or late in the afternoon, while shorter shadows indicate the image was taken closer to mid-day.

As the sun rises and sets in the same place every day, we can use shadows as a rudimentary compass. Long shadows in an image will be pointing in either a westerly or easterly direction, while short shadows will be pointing in either a northerly (in the Northern Hemisphere) or southerly direction (in the Southern Hemisphere).

Moss

Moss tends to grow in damp and shady areas, and therefore is more prominent on the side of trees and walls that faces away from the direction of sunlight. Therefore, in images taken in the Northern Hemisphere, the presence of moss on a tree indicates that is the northern side of the tree and therefore the image was likely taken facing a southerly direction, and visa versa for the Southern Hemisphere.

Rivers

If you know the name of the river that is present in an image, and can work out which direction the water is flowing, it is possible to work out which side of the river the image was taken from. This is much easier when working with video imagery than still photography.

Step 5: Triangulate the image

If you can recognise landmarks in the image, either natural or man-made, you can use this to work out exactly where the camera was when the image was captured. The more landmarks you can recognise, the easier this will be.

In the example above, we have identified a castle (red), a turret on a bridge (green), a small turret (blue), and a larger turret (yellow). The bridge is in front of the castle, and slightly to the left, while the smaller turret is in front of the larger turret, and also slightly to the left.

We can draw one line from the castle passing by the right edge of the turret on the bridge, and another line from the larger turret passing by the right edge of the smaller turret.

Where the lines eventually intersect will give us an accurate estimate of where the image was taken from. In this case, another turret that is jutting out into the river.

Final thoughts

The majority of the tips in this article will only give you evidence that an image was taken in a certain location, not proof. Therefore, aim to find as many different pieces of evidence as you can to give you the best chances of successfully geolocating the image.

Also, be aware that it’s highly likely you won’t find any evidence at all - people only ever post their success stories in things like this- so don’t be disheartened if you are’t successful.

If you want to practice your geolocation skills, have a look at GeoGuesser. This is a free game which provides you with a selection of images and challenges you to work out where in the world they were captured.

If you want to read some really in depth case studies of geolocation, I would suggest visiting the Bellingcat website. This has some incredible examples of geolocation being used to save lives and catch criminals with really detailed methodologies.

Found this interesting? Follow me on LinkedIn for more intelligence, OSINT, and geopolitics content

Subscribe to keep reading

This content is free, but you must be subscribed to OSINTSUM to continue reading.

Already a subscriber?Sign In.Not now

Join the conversation

or to participate.