CAM-911
Something new in front
A neighborhood street at midday, seen from a camera on a light pole. Further up, the produce market with its striped awning, the corner bar and two cars on the road. Someone is walking along the right-hand sidewalk. And across the bottom left of the frame, where yesterday you could see the crossing and the store doorways, there’s now a site fence: scaffold tubes and green mesh, put up from one day to the next. The camera is still recording, and what it records is the mesh. IRIS sees that part of the frame it used to see is now covered, and since when. It doesn’t know who put the construction fence there or why, and it doesn’t assume.
Turn on IRIS Vision
AI-generated image
What IRIS understands
What it recognizes in this scene
- new construction fence in front
The rule
You decide once. IRIS applies it always.
If
part of the frame that used to be visible has been covered by something that was not there for longer than the time set
Then
- alert maintenance with the camera, the time it appeared and the video from before and after
- keep the reference picture and the current one side by side, so you can see which piece has been lost
- flag that camera as partly covered while it stays that way, so nobody counts on that piece of the picture
The instruction that set it up«Tell me if something has appeared in front of a camera that was not there before and is still there the next day.»
What changes
What changes for whoever answers for the cameras working
There’s almost never any bad intent. Someone puts up scaffolding to repair a facade, a store wheels out a board, a truck parks where it can. Nobody tells the control room, because it doesn’t occur to anyone that this has anything to do with the cameras. On screen the camera stays green: in focus, streaming, recording. And three weeks go by. The day somebody needs the video of that sidewalk, what there’s is green mesh. With IRIS the alert arrives on the day the construction fence goes up, with the earlier picture next to the current one. Then you decide: move the camera, change the field of view, or simply know that this stretch of street isn’t covered this month. What you can’t afford is finding out late.
What people usually ask
Will it alert me every time a truck parks in front?
Only if you want it to. You write the rule, and the first thing you set is how long something has to sit in front of the lens before it’s worth an alert. A delivery truck for twenty minutes, usually not. A construction fence still there the next morning, yes. That line isn’t drawn by IRIS: it is drawn by whoever knows that street and knows what’s normal on it. And it can be changed the day it stops being normal.
How does IRIS know that it wasn’t there before?
Because it compares against what that same camera has been seeing. Every camera has its reference picture, taken on an ordinary day with nothing odd in it, and IRIS checks whether part of the frame has stopped matching it and for how long. You don’t have to describe the street or draw zones by hand: you just ask it to warn you when something appears in front and doesn’t go away. If the frame goes back to normal, the alert closes on its own.
Does IRIS say whether somebody put it there on purpose?
No, and it never will. IRIS states a fact: there’s something in front of the lens that wasn’t there before, on this camera, since this time. It may be authorized scaffolding, a bag that blew there or a board somebody hung up without thinking. Who put it there, and with what intention, is decided by a person looking at the video, not by the machine. That line is drawn on purpose, and crossing it would mean selling something we can’t demonstrate.
See it for yourself
A new case each time. Always the same screen.
It plays by itself and shows one case from start to finish. It’s 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
What happens in the small hoursdoesn’t have to wait for the morning report.
IRIS understands it there and then and tells whoever is on call, with the camera and the time. The next day there is nothing to reconstruct: it’s already written down.
Tell us your problem and we’ll say whether IRIS understands it, or not yet.
A person answers, the same working day.