Solutions by typeInfrastructure integrity
Who looks at the slope between one inspection and the next?
A slope warns slowly, and nobody watches slowly
Nine forty, a clear morning over the tailings pond. The basin is nearly full and the mud comes up close to the crest. Water is seeping out halfway down the outer slope and running away, cutting a catch basin between the earth and the stone. Yesterday that catch basin was a thread. Nobody knows that for certain, because nobody has yesterday’s photograph taken from the same spot with the same framing to hold up beside it.
9 cameras across 5 sectors
This is how you ask for it
«Tell me if water appears on the outer slope, if the catch basin grows or if the mud creeps toward the crest, and show me yesterday’s image beside it.»
By the time a slope really moves, there’s no margin left. The material comes down onto the haul road, onto the next lane or onto the working floor, and at that point the only thing left to do is count who was underneath. Before that there were weeks of warnings: a damp patch, a catch basin that grows a hand’s width, a crest that has lost its line. They’re all in photographs somebody took, filed in different folders and never once placed side by side.
You decide once
You decide this once. You say which slopes are watched, where the crest line sits in the image, how much something has to change before it counts, how often the comparison runs and who gets told: if water appears on the outer slope or the catch basin widens against last week, alert the person responsible for the pond with both images. IRIS checks it every day, including the days with no inspection scheduled, and above all through the two weeks when the team is busy elsewhere.
A slope almost never fails all at once. It fails after warning for days in a language nobody reads: a damp patch halfway down, a catch basin widening by a hand’s width, a crest that has lost its straight line, a bulge where there used to be a plane. The problem isn’t missing information. The problem is that seeing it requires yesterday’s photograph, taken from the same place with the same framing, and nobody has that photograph.
And this isn’t only mining. At twenty past six in the evening, a stack of recycling bales standing well above the wall and starting to come down onto the tipping floor. A slurry pond on an urban site, a hand’s width from the rim, yellow railings alongside. A road cutting that has shed earth and stone across the right-hand lane while a car approaches that hasn’t seen it yet. They’re all the same sentence: today this isn’t how it was.
IRIS keeps the framing still and compares the scene with itself. When something departs from what it was — water where there was none, material that has moved, a stack over the height of the wall, mud creeping toward the crest — it says so and shows both images together, today’s and the earlier one. It doesn’t measure the inch and it doesn’t predict the failure: it puts the change nobody had time to compare in front of the right person.
What changes
Before, there was an inspection every so often and a folder of photographs taken from roughly similar spots, impossible to line up against each other. Between two inspections, the slope was on its own. Now the same camera watches the same frame every day, and when something changes the comparison arrives already made: today’s image, the earlier one and the area marked. The walk-through doesn’t disappear; it arrives knowing where to look first.
Where we have seen it
CAM-380A loaded truck leaves by a way that isn’t a way out
«Tell me if a loaded truck leaves the stockpile by the lower track instead of going over the truck scale.»
CAM-433The bale stack has slid onto the yard
«Tell me if a stack goes above the wall or if material slides onto the yard.»
CAM-65Landslide onto the road
«Tell me if material comes down from the slope onto the road, and show me if the slope looks different from the last inspection.»
CAM-413No hard hat, no safety vest, right at the edge of the open tank
«Tell me if anyone works at the edge of a basin without a hard hat or a vest.»
CAM-58The slurry pit close to the brim
«Tell me when the slurry gets within a hand’s width of the top of the pit, before it goes over and I find it all over the site.»
CAM-383A bucket of gray gravel on top of the wrong stockpile
«Tell me if a loader drops material on a stockpile that is not that pile’s material.»
CAM-407Half the basin is moving and the other half isn’t
«Tell me if the surface of a basin stops moving while the system says the machine is running.»
CAM-396Seepage cutting a catch basin down the tailings dam face
«Warn me if water appears on the outer face, if the gully grows or if the slurry gets close to the crest, and show me yesterday’s image next to it.»
What it doesn’t do
IRIS does not measure displacement and does not replace instrumentation. Piezometers, inclinometers and survey work measure what a camera cannot, and they are what governs the safety of a pond.
Nor does it say whether a slope is going to fail. It says that today it looks different from yesterday, and where. And it cannot see below ground, which is where almost every real failure begins.
Questions
- Does this replace the instrumentation on the pond?
- No, and that matters. Piezometers, inclinometers and survey work measure things a camera can’t, and they’re what governs the safety of a pond. What IRIS adds is continuity: it looks at the surface every day, including the days with no inspection due, and speaks up when water appears where there was none or when something has moved. It’s the eye between six in the evening and seven in the morning, when there’s nobody up there.
- How does it know the catch basin has grown?
- Because it compares the same camera with itself. The framing doesn’t move, so today’s image and last week’s can be laid over one another and the mismatch shows. IRIS marks the area that has changed and shows you both. It doesn’t give you inches: it gives you the place and the date it started. The inch remains the surveyor’s job, and that’s exactly the line between what can be claimed here and what can’t.
- What about at night, or in dust and haze?
- At night you see little, and with dust in the air you see less. We aren’t going to dress that up. For this word it matters less than for others, because a slope doesn’t change in an hour: the useful comparison is one day against another, and one stretch of good light a day is enough. What you do have to do when setting it up is pick that stretch and always compare at the same hour, so a shadow doesn’t look like a new catch basin.
- Does it work for stockpiles and road cuttings, or only for ponds?
- It works for anything that is a mass of material with a shape that should hold. A tailings pond, a stack of bales, a gravel pile, a road cutting, a slurry pond on a jobsite. The question is always the same: is this the way it was? What changes from one place to another are the marks you paint on the image and who receives the alert: on a site it is the safety manager, on a road it is the control room, because what lies underneath is a lane with cars in it.
The real interface, step by step
You set the rule once. IRIS applies it every time.
You will see a summary of the interface, played step by step and hands-free. Each round starts with another case. The complete tool doesn’t fit in a demo.
- The rule gets built
- It goes to every camera
- Published, it watches alone
- And then it fires
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.
And the best part
“The white van, Tuesday afternoon.”That’s enough to find it.
You type it the way you’d tell a colleague, and the matching minutes come back, each with its camera and its time. You don’t need to remember the exact moment.
Tell us your problem and we’ll say whether IRIS understands it, or not yet.
A person answers, the same working day.