GeoScout
Image geolocation through street-level comparison. Upload a photo, draw a search area, and find exactly where it was taken. Four imagery sources, four comparison algorithms, zero guesswork.
Overview
GeoScout is a geolocation tool for finding where a photo was taken. You upload a reference image, whether it's a street-level photo, a screenshot from a conversation, or something pulled off social media, then draw a search area on an interactive map, and it does the rest.
It pulls existing street-level imagery from across that area and runs four comparison algorithms against your reference photo: perceptual hashing for structural similarity, SSIM for luminance and contrast, color histogram matching, and template matching. Each gets a weighted score. Anything above your threshold shows up as a pin on the map with a direct link to view the matched image at that exact location and heading.
Built for investigators, journalists, and researchers who need to verify locations from imagery. Available as a direct download, on GitHub, or live as a Tor hidden service.
Search Modes
Crowdsourced street-level imagery with 2+ billion photos. Free token, no cost. Best coverage in Western cities and Europe. The default mode for most investigations.
Highest quality imagery available. Two-pass scan — coarse grid first, then fine grid around hits for building-level precision. $7 per 1,000 images with $200/month free credit.
Best coverage of Chinese cities. Handles WGS84 → BD09 coordinate conversion automatically. Requires a Baidu developer key.
Overhead imagery comparison for aerial and satellite photos. Free through ESRI and Sentinel-2. Configurable zoom levels from region down to individual buildings.
Features
Screenshots
