CAM-933
Bands across the picture
A seafront avenue running into the mouth of a tunnel. There are palms on both sides, a promenade with lamp posts and bollards, water at the top left, a white van at the bottom, a dark car in the middle and a small white hatchback heading into the tunnel, where the lights of the cars inside are already visible. Over all of it lies a weave of fine horizontal lines that covers the whole frame, and the colour is dulled, washed out. The lines hide nothing completely: they dirty everything a little. IRIS sees a fixed pattern that is not part of the scene. Where it comes from, it does not know.
Turn on IRIS Vision
AI-generated image
What IRIS understands
What it recognises in this scene
- fixed lines across the picture
The rule
You decide once. IRIS applies it always.
If
a fixed line pattern appears that is not part of the scene and stays put while everything else moves
Then
- alert maintenance with the camera, the time and the video
- put the healthy image of the same view alongside so the difference is visible
- note whether it is there always, on and off, or only at certain hours
The instruction that set it up«Tell me if a line pattern appears on a camera that was not there before and does not go away.»
What changes
What changes for whoever answers for the cameras working
With banding you can still see that a car goes by. What is lost is the fine stuff: the plate at the tunnel entrance, the true colour of a van, a face at middle distance. And the search is lost with it: asking for a white van over a washed-out picture returns worse results, and nobody knows why. What usually happens is that somebody goes up to clean the dome and comes back down with exactly the same picture, because the glass was clean. The IRIS alert says this is a line pattern and not dirt, and at what hours it appears. That makes the trip worth taking, and the hour is often half the answer.
What people usually ask
Does cleaning the glass fix it?
No. The bands are not on the glass, they are in the picture. They usually come from electrical interference near the cable or from the sensor itself, which is why somebody can go up, clean the dome and come back down with the same image. The IRIS alert tells the two apart: it says there is a line pattern, not that there is dirt. And it adds at what hours it appears, which is often the clue that leads to the cause: if it only happens when something nearby switches on, you know where to look.
Is a picture with banding any use at all?
For quite a lot, and not for everything. You can still see that a vehicle goes past, where and at what time, and it still works for counting. What suffers is fine detail and colour, and with colour goes the search: if white is no longer white, searching by colour returns worse results. It is good for seeing, not for reading. That is the distinction worth being clear about before you need the video, rather than on the day you need it.
Will IRIS not mistake banding for a shutter or a fence?
That is the right question, and the difference is in the behaviour, not in the shape. A shutter or a fence occupies the part of the frame where it stands, gets covered when something passes in front of it and changes with the daylight. The defect pattern is over the whole frame, stays there even when a lorry goes past, and does not change shape. That is what gets compared. Even so, an alert is an alert and not a diagnosis: a person confirms it by watching the video, which comes attached.
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
What does a shopping centre see beyond the visitor count?
12 cases: 4 raise an alert when something happens and 8 only measure. You see each image as it is, and then with what IRIS understands on top.
What does an airport see before the queue forms?
14 cases: 7 raise an alert when something happens and 7 only measure. You see each image as it is, and then with what IRIS understands on top.
And the best part
The door was left open at two.Finding out at nine is no use to anyone.
The video tells you at nine. IRIS says it at two, with the camera, the time and the images from that moment. The difference between the two is whether someone can go.
Tell us your problem and we will say whether IRIS understands it, or not yet.
We reply the same working day.