Journalist Photo Verification
Upload the photo in question. Verify before you publish.
A viral photo claims a location. Your job is to confirm or refute it before the story runs. Geolocation is the step that converts a claim into a verifiable place on a map.
A viral photo, a different city
The claim
A photo circulated on social media claiming to show a protest at Place de la République in Paris. The image showed a wide boulevard, a crowd, and a distinctive building corner.
Oceanir result
Coordinates: 50.8467, 4.3525. Brussels, Belgium. Confidence: 82%. The building corner matched the Bourse de Bruxelles, not any structure at Place de la République. The discrepancy was documented in the published correction.
Verification checklist
Screenshot or print this section. Work through it for every photo before publication.
- 01
Source the image
Who shared it, when, with what claim. Record the URL, the timestamp, the platform, and the exact wording of the claim being made. Screenshot the original post.
- 02
Reverse search
Run the image through TinEye and Google Images. If it appears on a stock photo site, an older news story, or a different geography, that's a finding.
- 03
Geolocate
Upload to Oceanir (above). Get ranked candidates with coordinates and confidence. Note the visual cues that drove each candidate.
- 04
Cross-check Street View
On D3 (Pro), pull Street View imagery at the candidate coordinates and compare scene geometry. A building corner, a road sign, a tree line all confirm or refute the match.
- 05
Check shadow angles
Use the timestamp and the candidate latitude to compute the expected sun elevation and azimuth. Compare to shadows in the image. A mismatch means wrong time, wrong hemisphere, or both.
- 06
Verify with secondary sources
Local news, weather records, satellite imagery, on-the-ground contacts. A single geolocation result is a lead, not a confirmation. Corroborate before publishing.
- 07
Document your methodology
Write down every step you took, every tool you used, every result you got, and every judgment call you made. If you can't reproduce your reasoning, neither can your editor or your reader.
How to cite Oceanir evidence in your story
Disclose the tool, the confidence score, and any cross-checks you performed. Reader trust depends on transparent methodology.
"Analysis conducted using Oceanir visual geolocation AI (oceanir.ai) returned a 87% confidence match for [location]. The result was cross-checked against Street View imagery."
Replace the percentage and location with your actual result. If you ran a D3 visual verification, add: "A D3 visual verification pass confirmed the scene geometry against Street View at the candidate coordinates."
Verify another photo
Free surface scan. No signup. Coordinates and confidence in seconds.
Frequently asked questions
How do journalists verify a photo before publishing?+
Source the image, reverse search it, geolocate it, cross-check Street View, verify shadow angles against the claimed time, corroborate with secondary sources, and document the methodology. Oceanir handles the geolocation step and produces an exportable evidence bundle.
Can I cite Oceanir in a published story?+
Yes. Use language like: "Analysis conducted using Oceanir visual geolocation AI (oceanir.ai) returned a 87% confidence match for [location]. The result was cross-checked against Street View imagery." Disclose the tool, the confidence, and the cross-check.
What confidence score is publishable?+
There's no universal threshold. A 90%+ match with distinctive visual cues (a unique landmark, a readable sign) is strong. A 60% match on a generic suburban street needs more corroboration. Disclose the score and let the reader weigh it.
Does Oceanir work on photos from social media?+
Yes. Social platforms strip EXIF and GPS on upload, and Oceanir does not rely on metadata. It reads pixels: architecture, signage, road markings, vegetation. Screenshots from Twitter, Facebook, Telegram, and WhatsApp all work.
How do I document my geolocation methodology?+
Record the source image, the Oceanir candidates with coordinates and confidence, the visual cues flagged, the Street View comparison result, the shadow angle check, and the secondary sources consulted. Oceanir's export bundle captures the first four; you add the rest.