CAM-923
Flat, no blacks
A wide crossroads with palm trees, mid-morning. Everything is there and can be made out: two white vans stopped at the far side, a blue car entering the intersection, a pale car on the crossing, people standing on both sidewalks, the awnings along the façade. And still the picture looks as if a gauze had been laid over it. The shadows aren’t black, they’re gray. The whites never reach white. The colors are washed out and nothing separates from the background. IRIS sees that the distance between the darkest and the lightest parts of this frame has narrowed and that the scene has lost its separation. It doesn’t know whether it’s fog, a dome scratched by years of sun, damp inside the housing, or a setting somebody moved.
Turn on IRIS Vision
AI-generated image
What IRIS understands
What it recognizes in this scene
- frame with no blacks
The rule
You decide once. IRIS applies it always.
If
the range between the dark and light areas of the frame narrows against that same camera’s own range and does not come back on its own
Then
- alert with the camera, the time and the video of the day the picture began to flatten
- place the reference picture beside today’s, so the difference is seen rather than debated
- say in the alert whether it lasts a while, like fog or condensation, or whether it has been days without coming back
The instruction that set it up«Tell me if a camera loses contrast compared with how it used to see, and put the earlier picture next to today’s.»
What changes
What changes for whoever answers for the cameras working
This is the fault nobody reports. Nobody calls the service desk to say the picture looks “a bit washed out”: you can still see, and that’s enough to do nothing. So it stays for months. The cost turns up the day somebody has to search: a dark gray sweater against gray asphalt, a backpack leaning on a wall of the same tone, a person standing beside a white van. All of it was in the picture, and nobody saw it, neither the person nor the machine. IRIS compares this frame with the one this same camera used to give and raises an alert when the separation drops, with both pictures side by side. The difference becomes visible; it stops being an opinion.
What people usually ask
Does fog count as a flat image? We get fog thirty mornings a year here.
Fog flattens the picture in exactly the same way, which is why the rule looks at two things: how long it lasts and whether it clears by itself. A foggy morning lifts by midday; a scratched dome doesn’t. You can ask for fog not to raise a fault alert, while whatever doesn’t clear still does. And even so, knowing that this spot has been blind since six in the morning is useful: it isn’t a fault, it’s operational information, and whoever decides where to send a patrol will take it.
It looks fine to me on the monitor. Isn’t this splitting hairs?
On a big monitor, with the scene in front of you and knowing what you’re looking at, nearly everything looks fine. The problem shows up in the other two situations: when you have to find something in four hours of recording, and when that scene is being watched alongside eleven others. There a picture with no blacks hides things that used to be obvious. And IRIS doesn’t compare with an ideal or with another camera: it compares with what this same camera gave. If nothing has changed, nothing is raised.
Can’t you just turn the contrast up on the recorder?
Turning the contrast up afterward doesn’t bring back what was never recorded: it spreads the same gray over more gray, and adds noise along the way. If the veil comes from a scratched dome, from damp inside the housing or from a branch filtering the light, the fix is at the camera, not at the monitor. All you need first is to know which of the four hundred is like this and since when, and that’s exactly what IRIS provides. The repair is then done by a person.
Demo · IRIS at work
Everything comes in. Only what matters gets through.
This sums up what IRIS does with one alert, from the moment it appears until it’s closed. The real screen holds far more; this is the main thread.
- The hour, as a grid
- You look where it’s dark
- You jump to that minute
Next step
When does a bad review get decided?
15 cases: 11 raise an alert when something happens and 4 only measure. You see each image as it is, and then with what IRIS understands on top.
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.
And the best part
There’s a moment when you can still act.It’s while it’s happening, and it’s short.
After that, all you can do is tell the story. IRIS speaks up at that moment — camera, time and images — and whoever is on call decides what to do with it.
Tell us your problem and we’ll say whether IRIS understands it, or not yet.
A person answers, the same working day.