CAM-934
Dots that never move
A square with a big palm tree in the middle, seen from some height. On the right there’s a café patio with cream umbrellas; on the left, an avenue with parked cars and two crosswalks; in between, people crossing and a cart with a striped awning. And there are two things that don’t belong to the square: a thick black line, slightly tilted, running from top to bottom down the left third and cutting across the road and the crossing; and small black squares scattered right across the frame, on the facade, on the asphalt, beside the palm. IRIS sees that those marks stay in the same place whatever passes in front of them. What’s underneath, nobody knows.
Turn on IRIS Vision
AI-generated image
What IRIS understands
What it recognizes in this scene
- always in the same spot
The rule
You decide once. IRIS applies it always.
If
there are marks that stay at the same point of the frame whatever passes in front
Then
- alert maintenance with the camera, the time and the video
- say where in the frame they fall and how much picture they cover
- look again every week and report if they are spreading
The instruction that set it up«Tell me if there are dots or lines that always stay in the same place in the picture.»
What changes
What changes for whoever answers for the cameras working
Somebody walks across that crossing and, for five feet, disappears. In the recording there’s nobody. Nobody erased them: they never arrived, because the piece of sensor that should have seen them doesn’t register anything. A small dot is harmless until it lands exactly on what the camera was there to see: a doorway, a register, a lane. And the human eye gets used to those marks in two days and stops noticing them. IRIS doesn’t get used to them: it counts them, says where in the frame they fall and warns when they spread. This is the fault you can plan for, and planning costs far less than discovering it on the bad day.
What people usually ask
Can this not be fixed in software?
No. The sensor is the part that collects the light; where it is broken nothing arrives, and no filter can invent what was never recorded. It can be masked on screen, and masking it is worse: the mark disappears from the monitor and is still missing from the archive. The camera gets repaired or replaced. What IRIS does is say that it’s there, where it falls and how much picture it takes away, so the replacement can be planned instead of discovered on the day the video is needed.
It’s four little dots. Does it really matter?
It depends entirely on where they fall. A dot over the sky will never bother anybody. The same dot over a strongroom door, over a license plate or over a crosswalk hides exactly what the camera was put there to watch. That’s why the alert doesn’t just say there are dots: it says which part of the frame they occupy and how much picture they take away. With that, you decide whether it’s urgent or whether it waits for the next maintenance round.
How does it tell a dead pixel from a fly or a smear on the glass?
By how it behaves over time. A fly moves or eventually leaves. A droplet dries and changes shape. A smear on the glass looks different depending on where the light comes from. A sensor dot is identical at midday and at three in the morning, in sunshine and in rain, and doesn’t shift by a single pixel in weeks. That consistency is what IRIS looks at. And even so, the person with the video in front of them is the one who decides whether anybody goes up the pole.
A different case every time
Watch what happens. Then watch what IRIS does with it.
Not a recorded video and not a drawing: it’s a summary of the real tool, with a different case each time round. The whole tool doesn’t fit in a box.
- The hour, as a grid
- You look where it’s dark
- You jump to that minute
Next step
Who watches a whole station at rush hour?
17 cases: 11 raise an alert when something happens and 6 only measure. You see each image as it is, and then with what IRIS understands on top.
Who sees the sales that never reach the register?
10 cases: 4 raise an alert when something happens and 6 only measure. You see each image as it is, and then with what IRIS understands on top.
And the best part
You don’t need to buy cameras:you need yours to understand what they see.
Nothing gets built, mounted or rewired. It plugs into what is already up and running, and the only thing that changes is what those cameras understand.
Tell us your problem and we’ll say whether IRIS understands it, or not yet.
A person answers, the same working day.