OSINT Photo Geolocation

Professional geolocation follows a repeatable framework: preserve the original file, extract all available metadata, analyze visual clues, generate hypotheses, verify independently, and document the chain of custody.

Amateur mistakes include starting with reverse image search (finds copies, not coordinates), trusting a single AI answer without verification, or uploading sensitive evidence to untrusted cloud parsers that retain files indefinitely.

whereisthis.place maps to stages one through three of the framework. Stages four through six remain human-driven — as they should be for any finding intended for publication or legal use.

Every OSINT geolocation case should begin with EXIF. The operation takes seconds and costs nothing. GPS coordinates, when present, reduce a global search to a map pin.

Beyond GPS, EXIF yields DateTimeOriginal (event timing), Make/Model (device capabilities — drone photos imply aerial perspective), and FocalLength (field of view estimation for 3D scene matching).

Client-side parsing on whereisthis.place ensures source photos are not uploaded during this stage — critical when handling whistleblower material or pre-publication journalism evidence.

When EXIF is empty, OSINT analysts decompose the image into clue panels: architecture, vegetation, infrastructure, text, vehicles, and shadows. Each panel narrows the geographic search space.

AI geolocation automates clue weighting at scale — evaluating thousands of micro-features humans might miss. whereisthis.place returns five ranked hypotheses rather than one answer, preserving the analyst's responsibility to verify.

Manual techniques still matter for low-confidence AI output. Bellingcat-style workflows isolate signage for OCR, match roof tile patterns to regional catalogs, and use SunCalc with EXIF timestamps to constrain viewing direction.

Geolocation OSINT is as much about debunking false claims as confirming true ones. A viral photo 'from Gaza' with EXIF GPS in suburban Ohio is an immediate debunk. A photo without metadata requires harder work.

Verification hierarchy: (1) EXIF GPS matching visible terrain, (2) Street View structural match at AI top prediction, (3) independent corroboration via local news or weather data, (4) expert community review.

Never publish coordinates based on AI alone. Minimum standard for journalism: two independent verification methods plus editor review. See our acceptable use policy for ethical constraints.

A user submits a photo claimed to show a protest in City A. Step one: EXIF scan. GPS tags place the capture at 51.5074° N, 0.1278° W — central London, not City A. Case debunked in seconds if the claim required a different country.

Alternative scenario: EXIF stripped, AI ranks London 68%, Manchester 14%, Birmingham 9%. Analyst crops the visible Underground roundel and red double-decker bus — strong London indicators. Street View at the top coordinate shows matching brick facade and traffic light placement. Verified.

The interactive OSINT workflow below walks through decision branches for similar cases — when to trust EXIF, when to escalate to AI, and when to go fully manual.

whereisthis.place handles EXIF + AI geolocation. Supplement with Google Earth Pro (historical satellite), SunCalc.net (shadow analysis), Overpass Turbo (OpenStreetMap feature queries), and InVID verification plugin (metadata + reverse search for video).

For team workflows, export EXIF dumps and AI prediction JSON into case management tools (Hunchly, Maltego) to maintain audit trails. Hash originals at intake; never edit before metadata extraction.

Not every geolocation case needs the full manual toolkit. Escalate deliberately — over-investing in a solved EXIF pin wastes time; under-investing in a 22% confidence AI spread risks publishing false locations.

Publishable OSINT geolocation depends as much on documentation as on coordinates. Editors, courts, and community reviewers ask how you knew — not just where. This workflow produces a defensible case file alongside whereisthis.place analysis.

Complete each step in order at intake. Retroactive documentation after publication invites challenge and is harder to reconstruct under deadline pressure.

Two images arrive in the same newsroom tip line. Image 1 claims to show damage in City X; EXIF GPS places capture in City Y, 400 km away, with matching DateTimeOriginal. Terrain check at the EXIF pin shows flat agricultural land — incompatible with the claimed urban blast scene. Verdict: debunked in under three minutes using metadata alone; AI never needed.

Image 2 has no EXIF — Telegram forward. AI ranks City X 58%, City Z 19%, City W 14%. Analyst isolates a regional highway shield and fuel-price sign format visible in frame. Street View at the top coordinate in City X shows the same guardrail design and billboard skeleton. A municipal traffic camera still from the same date confirms weather conditions. Verdict: location corroborated with two independent methods beyond AI.

Same tool, opposite outcomes: the OSINT workflow's value is knowing when to stop (Image 1) and when to escalate (Image 2). Document both paths in the case file so editors see why one image was rejected and the other cleared.