Inputs
Multiple streams in
- optical
- live aerial RGB
- thermal
- IR channel, day or night
- telemetry
- position, heading, altitude
Many feeds in. One persistent object out.
Praxis fuses multiple sensing streams and tracks each object as a single persistent identity, with state, memory, and relationships.
Persistent Object
Vehicle 7F-21
- state
- moving, 45 km/h, heading 128°
- history
- seen in zone A for 12 minutes
- links
- near person A-23, restricted boundary
Drone 03 · Lidar Active
Survey · Sector 04
Mapping
Every flight builds the map
Praxis turns each patrol into structured data. Buildings, zones, and tracked objects, captured as the drone flies.
The scene persists, even after the camera moves
Praxis remembers what it scans. The zones operators marked stay put even when the camera moves to a completely different position.
Today
Video only
- Scan
- watch the stream
- Recall
- guess what changed
- Decide
- partial context
Praxis
Operational scene
- Pin
- mark a place
- Persist
- holds through motion
- Trigger
- alerts and handoff
The feed is the interface
Operators place persistent logic onto the world through video, not flat drawings on pixels.
- 01
Pin
Mark a road, gate, door, or boundary.
- 02
Classify
Turn pixels into objects and places.
- 03
Attach
Bind rules to the real location.
- 04
Remember
Keep it intact as the frame moves.
- 05
Act
Alert, hand off, trigger workflow.
Examples
Example 1
Live aerial video, structured into operational intelligence.
- 01
Ingest
Aerial video, live.
- 02
Detect
Vehicles, people, equipment per frame.
- 03
Track
Persistent IDs. Vehicle 7F-21 stays the same.
- 04
Apply
Spatial rules from operator-drawn zones.
- 05
Surface
Events, not clips. "7F-21 entered zone A."
Example 2
An operator places persistent logic on the world through the feed.
- 01
Pin
Gate, road, door, boundary.
- 02
Classify
What’s there, in object terms.
- 03
Attach
"Alert if anything crosses this line."
- 04
Remember
Pin holds as the drone moves.
- 05
Act
Alert, handoff, trigger workflow.