Skip to main content
Back to blogMethodology

Satellite Imagery Analysis: How We Verify Geolocation Claims in Conflict Zones

February 28, 20268 min readIntelligence Team

The Geoverification Challenge

In an era of sophisticated information operations, verifying that an image or video actually depicts what and where it claims to is one of the most critical — and difficult — tasks in open-source intelligence. State and non-state actors routinely recycle old footage, manipulate metadata, and stage events to shape narratives. Our geoverification queue processes hundreds of items per week, applying a systematic eight-method framework to each claim.

Eight Methods of Geolocation Verification

Our verification framework applies the following methods in sequence, with each successful method increasing the confidence score:

  1. EXIF and metadata extraction — examining embedded camera data, GPS coordinates, timestamps, and device signatures. While metadata is easily stripped or spoofed, its presence (or conspicuous absence) provides a starting point.
  2. Shadow analysis — calculating sun position based on shadow angles and comparing against astronomical data for the claimed location and time. This is particularly effective for verifying outdoor imagery and can narrow a claimed time to within 30 minutes.
  3. Landmark matching — identifying buildings, terrain features, infrastructure, and signage visible in imagery and matching against satellite imagery and street-level photography databases.
  4. Vegetation and terrain analysis — comparing visible flora, soil types, and terrain characteristics against known geographic and seasonal patterns for the claimed region.
  5. Weather corroboration — cross-referencing visible weather conditions (cloud cover, precipitation, lighting) against historical meteorological data for the claimed location and time.
  6. Vehicle and equipment identification — cataloging visible military equipment, vehicle types, license plates, and uniforms against known inventories and regional deployment patterns.
  7. Acoustic analysis — for video content, analyzing ambient sound for language, dialects, call-to-prayer timing, traffic patterns, and weapon signatures.
  8. Multi-temporal satellite comparison — comparing claimed damage or activity against before-and-after satellite imagery to verify that changes occurred within the claimed timeframe.

Confidence Scoring

Not every claim requires all eight methods. Our confidence scoring system assigns a verification level based on how many independent methods confirm the claim. A score of "verified — high confidence" requires at least four independent methods confirming the location, with no contradictory evidence. We explicitly label items where verification is partial or inconclusive, because transparency about what we do not know is as important as confidence in what we do.