Skip to content
IRIS NEURALContact

Solutions by typeIntrusion detection

Do I hear about it while they’re on top, or once they’re in?

Somebody is on top of the fence right now

9:40 p.m. on a Friday in June, still light. A hoarded jobsite in the middle of town. At the corner of the wooden construction fence somebody’s halfway up: one foot on the boarding, both hands on the top edge. There’s nobody inside; the crane is parked, the materials stacked, the machines locked. It took eleven seconds to reach the top, and people are walking along that sidewalk without noticing.

12 cameras across 8 sectors

This is how you ask for it

«Tell me if someone climbs the fence or gets in where nobody gets in, at any hour, and show me the camera so I know where they came through.»

What they take is almost never the expensive part: copper, an angle grinder, the diesel out of the machine tank. The expensive part comes after. Monday starts with an incident report, a machine that won’t start and a lost day on site that nobody gets back. And there’s a consequence nobody talks about: if the person who climbed over is a teenager, inside there is an elevator shaft, an open trench and a site distribution board, and then the problem stops being theft.

You decide once

You decide this once, and there’s barely any nuance, because on a fence no schedule applies: nobody comes in here, ever. You say which stretches are perimeter, who gets told by day and who by night, and what you want kept: video from ten seconds before, the camera and the exact point. IRIS applies it to all twelve stretches alike, at nine on a Friday evening and at four on a Tuesday morning, and it never tires of watching a fence where nothing happens for months.

Of everything a camera can see, this is the one with the least room for debate. Nobody climbs a fence by mistake. The fence is already a decision, made and written on the ground: you don’t come through here. So there’s nothing to interpret and nobody’s intentions to wonder about; all that is needed is for somebody to find out, and to find out soon.

And “soon” is the whole story. Finding out once they’re inside is good for writing a report. Finding out while they’re on top of the fence, both hands on the top rail and one leg over, is good for somebody to make a call, turn on a light, speak over the PA or send a guard who’s four minutes away. Same camera, same video, same fact: the only thing that changes is how many seconds it takes to reach a human being.

IRIS watches the fence line and alerts at the moment of contact, not once the person is already walking about inside. The alert carries the camera, the point and video from a few seconds before, which is the part you need afterward: how they got there, through where, and whether they came alone. And all of it accumulates. After a few months, the list of climbs doesn’t tell you who comes in: it tells you where, and it’s almost always two of the twelve stretches. With that you fix a fence, which is still the only thing that actually stops anyone.

What changes

Before, the perimeter alarm went off at three in the morning with no image behind it. Somebody had to decide, half asleep, whether to send a car on a forty-minute drive to find a plastic bag caught in the mesh. Now the alert arrives with the image of the moment, and in twenty seconds you can see whether there’s a person on the fence or the wind has moved a tarpaulin.

Where we have seen it

  • The crest of a dam late in the afternoon. Somebody is going over the mesh gate that closes the way through: both hands on the top rail and one leg already on the other side. Beside it, the control hut with its green steel door, the red no-entry sign wired to the gate leaf and the barrier on the track. Nobody else is in frame. IRIS sees that someone has crossed a closed barrier; it doesn’t know who they’re and doesn’t try to find out.

    CAM-144Someone climbs the gate onto the dam crest

    «Tell me the moment anyone climbs over the fence onto the crest, at any hour, and tell me which camera and where they got in.»

    See this camera →Critical infrastructure

  • Mid-morning at the terminal’s outer fence. A person is up on the chain link, a few feet from the control booth. On the other side, on the patrol track, a guard is looking that way. Beyond the fence lies the yard: stacked containers, a truck on its way out and the ship berthed at the back. IRIS marks the exact point of the fence that was crossed, the time and the direction, and saves the video of the seconds before.

    CAM-440Somebody has come over the perimeter fence

    «Tell me if somebody climbs or crosses the fence, and show me where and at what time.»

    See this camera →Ports

  • The outer fence at dusk, with the patrol road and the wall on the right. A person in facility clothing is halfway up the mesh, gripping the razor wire at the top. Another is already on the outside, running through the scrub. On the inner road a patrol vehicle is parked and an officer is on foot, and at the corner of the wall stands the watchtower with another officer inside.

    CAM-105Someone on the perimeter fence

    «Tell me the moment anyone touches the perimeter fence, not once they are already over it.»

    See this camera →The prison

  • Eleven in the morning in the sorting hall. The mesh gate of the shredder enclosure is wide open and a worker in a yellow vest has just stepped through it: one foot is already inside and he is walking toward the machine. The shredder is still running and the conveyor keeps feeding the hopper. A moving-parts warning sign hangs on the guard panel. IRIS sees the guard open, the person inside and the machine running, all at once.

    CAM-431Someone has walked into the shredder enclosure

    «Tell me if someone enters the shredder enclosure while the machine is running.»

    See this camera →Industry

  • Late afternoon at a small drinking-water plant, up on a clifftop. The compound is closed off with mesh and wire, and the open basin sits in plain view inside. On the hillside, a person has come over the fence and is leaning on the inner rail. A worker is coming across from the gate. IRIS doesn’t say who they’re or why they came: it says a closed boundary has been crossed.

    CAM-408Somebody came over the fence and is among the tanks

    «Tell me the moment somebody crosses the fence where there is no gate, at any hour.»

    See this camera →Industry

  • Mid-morning at the coastal wind farm. The camera looks down on the tower. Out on the external platform, next to the ladder, a person is standing with no harness and nothing holding them. Below, inside the yellow and black fence, a second person walks toward the tower door, which is open. There’s no vehicle on the access track. IRIS sees two people where, at this hour, nobody should be without notice.

    CAM-355Someone is on the turbine platform with no harness

    «Tell me if anyone is on a tower ladder or platform without a harness, or if someone enters the fence of a tower with no work scheduled today.»

    See this camera →Industry

  • Late afternoon at the venue’s access control. A dark van has been sitting on the crossing, in front of the barrier and right against the guard booth, tailgate toward the camera. On the right, on the restricted-area fence, a person is gripping the top rail with one foot on the wall. The guard on the sidewalk is looking the other way. Behind them, the public is arriving at the gates.

    CAM-488A van stopped at the barrier and someone climbing the fence

    «Tell me if anyone crosses the fence line, or if a vehicle stays stopped in front of the barrier for more than two minutes.»

    See this camera →Venues and trade shows

  • A fenced building site in the middle of the city, seen from above. At the corner of the timber construction fence a person is halfway up: one foot against the boards, both hands on the top edge. Inside you can see the debris dumpster, the churned-up ground and, further back, two workers in hard hats and safety vests. IRIS sees that a person is crossing the fence; it doesn’t know who they are.

    CAM-57Someone climbing the site fence

    «Tell me if someone climbs the fence or gets in where nobody is meant to, whatever time it is, and show me the camera so I know where they came through.»

    See this camera →The jobsite

  • A substation at the edge of the road, seen from a light pole on the sidewalk. Someone has stepped up onto the low wall and is holding on to the mesh with one foot on the rail, about to go over. Inside there’s nobody: the transformer, the control house and the closed gate. Here you don’t have to decide how many people are too many. One person inside the fence is already the incident.

    CAM-158Someone climbing the substation fence

    «Tell me the moment somebody touches the fence or gets over it, at any hour. Nobody is supposed to be in here.»

    See this camera →Critical infrastructure

  • Afternoon concert, front row. Someone vaults the separation barrier and lands on the stage side. A cup thrown from the stands is still in the air. Two security staff are already running over. IRIS does not know who they are: it does not recognize faces, does not compare them with anything and keeps nobody’s identity. What it sees is a movement — a body crossing a line nobody crosses — and an object flying where nothing flies. It detects a movement, not a person.

    CAM-490Someone has jumped the barrier toward the stage

    «Tell me the moment somebody crosses the barrier toward the stage or climbs a structure, and save me the video from ten seconds before.»

    See this camera →Venues and trade shows

What it doesn’t do

IRIS stops nobody and identifies nobody. It sees the climb, alerts and keeps the video; a person decides what happens next. And a camera is not a fence: if the boundary is broken, seeing the crossing more clearly does not fix that.

And it depends on the camera seeing. A dirty lens or a branch that has grown in front leaves a blind stretch, and that has to be known in advance, not the day something happens.

Questions

Would a detection system on the fence itself not be better?
It depends, and the two often work well together. A system mounted on the fence notices that something touched it, and that’s fast and cheap on a long perimeter. What it doesn’t know is what touched it: the wind, a dog, a branch or a person. That’s why many sites end up with the alarm turned off, which is the worst way to protect a perimeter. With a camera you do see what it was, and the person receiving the alert decides with the image in front of them instead of blind.
Does it alert while they’re climbing or once they’re in?
While they climb, if you set it that way, and that’s the recommendation. The rule is tied to contact with the fence, not to being inside, because the whole difference between a useful alert and a report lives in those seconds. The alert comes with video from a few seconds earlier, so whoever receives it sees the full scene: how they arrived, whether they came with anyone and where they’re heading. That’s what you use to decide whether to call, turn on a light or send somebody who’s nearby.
Does it say which point of the fence they came through?
Yes, and in the long run that is the most useful part of all. The alert carries the camera and the stretch, so whoever goes knows where to go instead of walking half a mile of fence with a flashlight. And because everything’s logged, after a few months you have a map: two stretches account for nearly all the climbs, almost always where there’s a dumpster parked outside, a bank of earth or a tree. That isn’t fixed with more cameras: it’s fixed by moving the dumpster or closing that stretch.
What about animals, wind and plastic bags?
IRIS says what it sees, and that’s exactly what removes most of those alerts: it tells a person from a dog, a branch or a flapping tarpaulin, and when it does alert it sends the image so it can be checked in two seconds. We don’t promise nothing extra will ever fire, because a foggy night or a dirty lens spoils any camera. What does change is that nobody has to leave the house to find out what it was: you look, you close it and you carry on.

See all the questions

The real interface, step by step

A camera looks. IRIS understands what it sees.

An extract of the real screen, no voice-over and hands-free. Each round poses another situation. What sits behind it is a good deal bigger than what you see.

  1. It lands in the tray
  2. You open it
  3. You confirm or dismiss
IRIS NEURAL
30ES
IRISDirectoGrabacionesGISIncidencias3996SituaciónCasosLPRAutomatizacionesIRIS DATA
Ask a question or make a request…Send
IRIS · Infinity Neural

Not seeing yours here?

These words are the ones we have built, not the ones IRIS understands. The list falls short on purpose: you ask for it in a sentence, and sentences don’t run out. Tell us what is in front of your cameras and we’ll tell you whether it understands it, or not yet.

See all solutionsTell us what you see

And the best part

A recording gets watched when nothing can be done.An alert lands while there’s still time to act.

IRIS raises the alert while it’s happening and says where: the camera, the time and the images from that moment. With that, whoever is on duty is still in time to act.

Call Spain · +34 902 02 70 91Get in touch

Tell us your problem and we’ll say whether IRIS understands it, or not yet.

A person answers, the same working day.