Live drone position on the map
Continuous coordinates of the intercepted drone, updated as the feed comes in.
From intercepted videoto coordinates on a map.
We turn intercepted FPV video feeds into precise drone positions on the map in real time, using AI-based cross-view matching against satellite imagery. No telemetry. No GPS. No cooperation from the drone.
Built for the Ukrainian battlefield. Integrates with leading FPV signal interception hardware.
// AOI · GRID 35XKE3.3 GHzLAT 48.0234°N · LON 37.7891°ECONFIDENCE ±18 m · HDG 247°
VIDEO FEED · 3.3 GHz
// LOCK 87%// LOCK 88%// LOCK 86%// LOCK 89%
Detection isn't location.
Today's frontline detectors are a major advance. A soldier with a Chuyka or ZORKO can see, in real time, what the FPV drone hunting them is looking at. That has saved lives. But it answers only one question: a drone is here, somewhere within a few kilometers.
The questions that decide the next thirty seconds are different ones:
- Where exactly is the drone right now?
- Where is it heading — which position, which vehicle, which crew?
- Can a friendly interceptor get a fix on it in time?
Answering those requires turning the intercepted video into a map coordinate. That has been hard to do — until now.
Built on the published state of the art in GNSS-denied UAV visual localization, adapted to intercepted, degraded battlefield video.
Three layers. One continuous geolocation pipeline.
01Ingest
Ingest
The system accepts the live intercepted video feed from your existing FPV signal interception hardware. Compressed, low-bitrate, lossy analog feeds are handled in stride — frame stabilization, artifact suppression, and quality gating are part of the pipeline.
02Match
Match
Each usable frame is matched against pre-loaded, georeferenced satellite imagery of the area of operation using state-of-the-art cross-view feature matching. The matching is built on the same family of techniques used in the latest published research on UAV visual geolocalization, adapted to the realities of intercepted, degraded battlefield video.
03Track
Track
A monocular visual odometry layer estimates relative motion between matches, and a fusion filter combines both to produce a smooth, drift-corrected trajectory. The output is a live drone position on your map, with confidence bounds and a heading estimate.
The video the drone sends home is also a map of where it is. We just have to know how to read it.
What the system gives you.
Interceptor cueing
Provides a fix that a friendly counter-FPV drone team can act on, instead of guessing from signal strength.
Works under EW conditions
GNSS-independent end to end. Effective where jamming and spoofing make conventional positioning unreliable.
Pattern of life
Aggregated tracks build a picture of enemy launch points, ISR routes, and reconnaissance patterns over time.
Where the project is — and how to engage.
In active development. Core architecture and team composition are defined. We are scoping feasibility on representative intercepted footage and onboarding the core engineering team.
We are open to discussions with:
- Defense Forces units interested in early field evaluation.
- Brave1 cluster partners and NATO-aligned funding programs.
- Interception hardware manufacturers for integration partnerships.
- Foundations and grant programs supporting Ukrainian defense technology.
- Engineers in computer vision, photogrammetry, monocular SLAM, and real-time video processing.
Letter of Intent · 7th Rapid Reaction Corps · DShV ZSU
Founded by Roman Prystatskyy (CEO) and Ihor Korobka (COO) · Berlin / Munich / Kyiv
// ACQUIRING TARGET48.3217°N · 35.1058°E
Turn every intercepted feed into a fix on the map.