CAM-910
Dirty
The usual junction, seen through glass that has not been cleaned for weeks. The picture is not shaky and it is not dark: it is veiled. A layer of dried spatter covers the whole frame and thickens into a broad band across the middle and the lower part. The big things are still recognisable — two zebra crossings, the green pharmacy awning, the blue café awning with people at the tables, a white car, a red one, a blue van on the crossing — but everything arrives soft, flat and dull in colour. The marks do not move: the cars pass underneath and the marks stay exactly where they were. IRIS sees that this camera sees worse than it did a month ago. What it does not know is whether the dirt is on the outside or the inside of the glass.
Turn on IRIS Vision
AI-generated image
What IRIS understands
What it recognises in this scene
- dirt on the optics
The rule
You decide once. IRIS applies it always.
If
a camera loses sharpness and contrast steadily over several days and the veil does not move with the scene
Then
- flag it as a camera that sees worse and put it on the cleaning list
- attach today’s picture alongside the one from the last day it saw properly
- notify maintenance and check it again after the clean
The instruction that set it up«Every morning, tell me which cameras see worse than they did a month ago and which ones need a clean, sorted by how much they have dropped.»
What changes
What changes for whoever answers for the cameras working
Nobody cleans a camera because nobody knows which one is dirty. Dirt arrives slowly: every day the picture is slightly worse than yesterday, and no single day is bad enough for anyone to pick up the phone. A month later the camera is still working — it gives a picture, it records, it shows green on the wall — and it no longer does the job it was bought for. The day somebody needs to read a plate, or the colour of a scooter, it is not there. With IRIS each camera is compared with itself, and the alert arrives when the drop is clear and a pole and a cloth still fix it. Dirt is not a fault: it is a clean nobody has asked for, because nobody has seen it.
What people usually ask
How does IRIS know the camera is dirty rather than that there is fog?
Because dirt does not move and fog does. IRIS compares each camera with itself over time, not with an ideal picture: it looks at whether sharpness and contrast have dropped, and whether the loss stays in the same parts of the frame day after day, in sun, in rain and at night. Fog lifts in the morning; dried spatter is still there next week. When the doubt persists, IRIS raises it anyway and whoever looks at the picture confirms it: a one-minute check is cheaper than a month of veiled camera.
Is the video from a dirty camera worth anything?
It is enough to know that something happened, and not enough to know what. With veiled optics you still see shapes, movement and vehicles; you stop seeing the number plate, the exact colour of a jacket or who was in front. And the underlying problem is not just that the picture is worse: it is that nobody finds out it is worse until the day that video is needed, and by then there is no way to get it back. That is why cleaning a camera is a security task, not a cosmetic one.
And once the technician has cleaned it, do you have to tell IRIS?
There is no need. IRIS looks at the camera again and compares it with its earlier state: if sharpness and contrast have come back, the alert closes by itself and it is logged that the camera was cleaned that day. If they have not come back, it stays open, and that is useful information too: the dirt may be on the inside of the dome, the glass may be scratched, or the fault may be the focus rather than the grime. An alert that does not close is how you find out that the visit did not fix the problem.
Demo · IRIS at work
A new case each time. Always the same screen.
It plays by itself and shows one case from start to finish. It is a summary: the real screen has far more functions than fit in this box.
- The hour, as a grid
- You look where it's dark
- You jump to that minute
Next step
Who is watching a whole quay at three in the morning?
14 cases: 13 raise an alert when something happens and one only measures. You see each image as it is, and then with what IRIS understands on top.
Who is watching the cameras of a whole branch network?
13 cases: 12 raise an alert when something happens and one only measures. You see each image as it is, and then with what IRIS understands on top.
And the best part
Video tells you what happened.An alert reaches you while it is happening.
Reviewing the recording helps you explain. Hearing about it at the time helps you get there. Two different things, and until now you only had the first.
Tell us your problem and we will say whether IRIS understands it, or not yet.
We reply the same working day.