A missing bucket tooth, a steel cable in the load, an oversized boulder: small problems that become crusher damage, downtime and lost throughput.
Atlas Merlin detects them on-site, in real time, and alerts the people who can act.
“Working with Atlas Merlin has allowed our operations team to focus on production. We have saved hundreds of thousands of dollars every month thanks to stopping foreign objects entering our production plant.”
Every application below runs on the same platform: the same on-site hardware, the same software. You choose what to watch for; the system it runs on doesn't change.
A camera mounted on the excavator watches the bucket continuously. The moment a tooth is broken or missing, the operator receives an instant alert. No connectivity required.
One missing tooth left undetected can cause catastrophic damage to crushers and conveyors downstream. Atlas Merlin catches it at the source, before it becomes a crusher event.
Nothing like this exists anywhere else.
A camera watches the ground and bucket ahead of the excavator. When it spots something that shouldn't be there (a drill rod, an old tooth, a steel cable, buried debris), the operator is alerted instantly.
Today, foreign objects are only caught by sensors further downstream, often after damage has already begun. Atlas Merlin detects them at the point of excavation, before they enter the processing chain.
Every tooth on the bucket is monitored continuously, and its wear tracked over time. From that wear rate, Atlas Merlin projects when each tooth will need replacing: a deterministic view of tooth condition across the bucket, not a subjective call from someone eyeballing it between shifts.
Replace teeth on evidence, before one cracks and drops into the material, closing the gap before it becomes a missing-tooth alert or a foreign object in the crusher.
A camera positioned above a grizzly crusher monitors incoming material in real time, flagging oversized boulders before they cause a blockage or damage.
When connected to the crusher's e-stop system, intervention is automatic. The same system analyses fragmentation patterns, giving your blasting team the data to close the loop between blast design and processing outcomes.
A camera watches conveyors and transfer points for blockages, spillage and material build-up, flagging problems before they escalate into a belt tear or an unplanned stop. When connected to the conveyor's e-stop system, intervention is automatic.
Caught early, a belt tear is a short maintenance window. Missed, the belt can pull through and off the system entirely: a reinstall that takes weeks. Atlas Merlin spots it while it's still the small job.
A single camera on a haul road delivers more than you'd expect. Originally deployed to monitor pantograph compliance (identifying which trucks are connecting to the overhead electrical trolley lines), the system expanded naturally from there.
Trolley-assist usage tracking, vehicle counts by type, congestion and bottleneck detection, all from one camera. If trolley-assist usage drops, the right people learn why, with the footage to back it up. One camera, multiple operational insights.
Today, dust suppression runs on radio: a supervisor calls dispatch to say their zone is bad. Dispatch has no objective way to know which area is worst, or where the water truck is needed most. Cameras across the site change that.
Visibility and dust are monitored continuously in every zone. When conditions deteriorate, the right people are notified at once: dispatch to route the water truck, the area manager to confirm it's actioned, the shift GM to flag a production impact. One signal, routed to whoever needs it.
A portal is a pinch point. Heavy vehicles, light vehicles and people all pass through the same narrow opening between surface and underground, often with limited sightlines and constant two-way movement. A single camera watches it continuously, tracking what goes in, what comes out and what's happening in the box-cut around it.
Congestion, near-misses and unauthorised movements are surfaced in real time, and the right people are alerted before a near-miss becomes an incident. Objective visibility of one of the most safety-critical points on the site.
Many operational risks are visible long before they become incidents.
If a person can recognise it, there's a good chance we can train a model to detect it.
We work alongside your team to develop, test and validate new detectors in live environments, quickly and at low cost. Rather than a heavy upfront commitment, we prove the value on your site first, then scale what works.
Typical consequences include:
Across deployed sites, Atlas Merlin identifies an average of two foreign-object events per month that would otherwise go unnoticed — helping you stop missing teeth, drill rods and other debris before they reach the plant.
Your number will differ, and for some sites it's far higher. A single crusher with no redundancy can lose millions from one event alone.
Every site is different. Our applications work from day one, then we tune them to your exact machinery, layout and conditions. The result beats anything generic.
Your model keeps learning. We review and refine it continuously, so it stays sharp as your site and conditions change. A static model falls behind. Ours doesn't, and that takes no effort from you.
Every improvement is included. No upgrade fees, no “next version” to buy, no project to schedule. Your models simply stay current, so the system you run in two years is better than today's. At no extra cost.
Deploy cameras and edge hardware.
We optimise models to your operation.
Receive alerts and operational insights.
Designed from the ground up for remote, harsh, and connectivity-constrained environments, and tested in them.
A missing tooth, an oversized boulder, a haul road nobody's watching: every one of these is costing someone, somewhere, right now.
If it's costing you, that's the conversation we want to have.