2026
DeMaVision
Production intelligence from camera footage
The factory already had 47 security cameras. DeMaVision turns them into a production sensor: it sees when a line runs, where it stalls and when packing waits for product, then checks that against the ERP. No new hardware, and no footage ever leaves the building.
How it fits together
- Cameras47 existing cameras on three recorders
- Vision engineRuns locally, on our own server
- ERP cross-checkWhat the cameras see, matched with what the system says
- Line stateRunning, stalled, or starved for product
- Daily reportWhere the time went, and why
The problem
Machines on a production line log their own numbers. The spaces between them usually don’t, and that’s where the minutes go missing: a crew waiting for product, a line that paused and nobody wrote down why.
What I built
Instead of buying sensors, I used what was already on the walls. DeMaVision reads the existing recorders, runs vision locally on our own server and turns each zone into a simple state: running, stalled or starved. Then it checks that against what the ERP says should be happening.
Why local
These are cameras pointed at people at work. Processing on our own hardware means no footage leaves the building, and the system measures lines and zones, not individuals.
Spec sheet
- Built for
- Production at Agrafresh
- Extra hardware
- €0
- Stack
- Python, computer vision, Hikvision recorders
- Status
- Measuring