OSINT & Geolocation
Overview
Geolocation is the OSINT discipline of working out where an image was taken from the visual evidence inside it — the same skill open-source investigators use to verify footage and place photos. I train it competitively in GeoGuessr, where you're dropped into a street-view panorama with no metadata and have to reason your way to a precise location.
How I read a location
It's a deduction process, not a guess. The cues I stack up:
- Road & infrastructure: line markings, bollard shapes, guardrail types, utility-pole design, driving side.
- Language & script: signage alphabet, language, and even font conventions narrow the country fast.
- Environment: vegetation, soil colour, and biome place the latitude band; mountains vs. plains do the rest.
- Architecture: building styles, roofing, fence and window patterns are strong regional tells.
- Sun & shadow: shadow direction plus hemisphere gives a rough heading and latitude.
- Camera meta: the street-view generation/car, camera blur, and rig artifacts often pins down the specific country very accurately.
Best guesses
A few standout rounds, with the reasoning that got me there:
How it maps to security work
Geolocation is a core OSINT capability used in threat intelligence, investigations, and incident response — verifying where a photo or video originated, corroborating claims, and building context from open sources. The underlying habit — extracting maximum signal from minimal evidence and reasoning to a conclusion — is exactly what investigative security work rewards.