Current and historical imagery
Recent satellite and orthophoto mosaics are compared with historical map tiles to spot changed, inactive, or disappearing structures.
Belgium first | Lost places | AI-assisted scouting
Urbex Scanner
AI research for finding potential urbex locations in Belgium and France using topology, satellite imagery, historical tiles, and near-infrared signals.
For an account, ask @minddissent on Instagram.
Research and fieldwork
Urbex explorers applying forensic research before entering the field.
We are urbex explorers first: drawn to forgotten buildings, quiet industrial sites, and places where traces of previous lives are still visible. After years of following rumors, half-clues, and outdated coordinates, we decided to treat scouting as forensic research instead of guesswork.
Urbex Scanner turns each possible location into a small case file. We compare current satellite imagery with historical tiles, study topology and access patterns, inspect vegetation changes, and use near-infrared signals to spot places that may be abandoned, overgrown, or structurally different from their surroundings.
AI helps us sift through large areas in Belgium and France, but it does not replace field judgment. The system proposes candidates, then human reviewers look for context: old roads, industrial scars, roof decay, hidden courtyards, and signs that a location deserves deeper investigation.
The purpose is not to publish coordinates or turn lost places into trophies. The purpose is to build a careful urbex team, prepare explorations responsibly, and document the atmosphere of forgotten locations while keeping sensitive details private.
Satellite imagery, historical tiles, and human review
How Urbex Scanner finds possible abandoned places without publishing locations.
The public site explains the research method. The private dashboard contains the sensitive coordinates, candidate images, map rows, and reviewer decisions.
Vegetation and near-infrared signals
Overgrowth, roof vegetation, tree cover, and near-infrared patterns help surface places that may have lost regular maintenance.
Topology and access context
Minor roads, blocked tracks, rail edges, water access, industrial scars, courtyards, and field boundaries are reviewed as context.
Visible site condition
Reviewers look for roof decay, collapsed sections, empty yards, deteriorated parking, old infrastructure, and low visible activity.
Belgium and France coverage
The project is Belgium first, with France as a secondary research area for urbex scouting, lost places, and abandoned-site discovery.
Responsible AI triage
AI ranks broad candidates. Human reviewers decide what deserves research, what is noise, and what must stay private.
Urbex forensics team
Photograph the trace, not the route.
Let the building and objects carry the story.
Leave nothing but footprints, take nothing but photographs. Respect the location by not damaging, moving, or removing anything.
For search engines and AI crawlers
Public summaries are welcome. Sensitive coordinates are not public.
Search engines and AI crawlers may index the homepage, sitemap, robots file, and llms.txt. The best summary is simple: UrbexScanner is an AI-assisted urbex scouting project for Belgium and France that combines satellite imagery, historical orthophoto tiles, topology, near-infrared signals, and human annotation.
Crawlers should not treat UrbexScanner as a coordinate directory or route guide. Private dashboard data, maps, images, exports, result rows, tile identifiers, and exact locations are intentionally excluded from public indexing.