Case Study: LiDAR 3D Model of Mine Site

TL;DR: This LiDAR 3D model mine site case study covers a combined UAV and terrestrial laser scanning programme at a Goldfields gold operation near Leonora, Western Australia, where an ageing set of as-built drawings made it impossible to design a CIL processing-plant expansion with confidence. ISS captured the pit shell, ROM pad, run-of-mine infrastructure and the congested process plant as a single georeferenced point cloud, then delivered a registered 3D model accurate to 25 mm against control. The model let the client's engineers route new pipework, site a thickener and confirm crane access without a single field re-measure — removing the clash risk that had stalled the design.

Key takeaways

• A single LiDAR programme combining a DJI Matrice 350 RTK with a Zenmuse L2 (corridor and pit capture) and a Leica RTC360 terrestrial scanner (plant and structural detail) produced one continuous georeferenced model spanning roughly 60 ha of mine site, from open pit to process plant.
• The whole model was tied to GDA2020, MGA2020 Zone 51 on AHD heights so it sat directly on the operation's existing mine grid and design surfaces — the new plant could be designed in the same coordinate system the mine already used.
• Terrestrial scanning held the congested plant to a registration accuracy of 25 mm, fine enough to design tie-ins, pipe routes and structural steel against measured reality rather than 12-year-old drawings.
• Capturing the brownfield plant as a point cloud removed the need for crews to access live, elevated and confined areas with tape and total station — a meaningful working-at-height and isolation-risk reduction during a producing operation.
• The 3D model became a reusable digital asset: the same dataset later supported clash detection, a thickener foundation design and a tailings spigot survey, spreading the cost of one capture across multiple engineering jobs.

The operation and the challenge

The client operates a mid-tier open-pit gold mine in the Northern Goldfields of Western Australia, near Leonora, feeding a carbon-in-leach (CIL) processing plant. The mine had been producing for over a decade and was planning a plant expansion — a new thickener, additional leach capacity and the pipework and structural steel to connect it into the existing circuit. Like most brownfield expansions on a long-running site, the project ran straight into the same wall: nobody trusted the drawings.

The available as-built documentation was a patchwork. Some areas dated to original construction roughly 12 years earlier; others had been modified during shutdowns and never fully redrawn. Pipe runs had been re-routed, platforms extended, and equipment swapped without the drawing set keeping pace. For a design team trying to thread new infrastructure through a congested, live plant, that is a serious problem — every clash discovered in the field during construction is rework, and rework during a shutdown is the most expensive rework there is.

The conventional fix would have been a total station crew picking up critical points, supplemented by tape measures and photographs. On a producing CIL plant that approach is slow, incomplete and hazardous: the areas of interest are elevated, congested with live process lines, and often only accessible under isolation. A point-by-point survey also captures only what the surveyor thought to measure — and the whole problem here was that the design team did not yet know which measurements they would need. They needed the entire plant, not a list of points.

Why this needed LiDAR

A LiDAR 3D model addresses the brownfield problem at its root: instead of measuring the points you think you need, you capture everything, then extract whatever the design throws up later. That distinction is why laser scanning has become the default for plant expansion and retrofit work. Three characteristics made it the right method here:

• Completeness. A terrestrial scanner captures millions of points per setup, recording every pipe, valve, beam, cable tray and platform in line of sight. When an engineer later asks "is there clearance for a 600 mm header along the eastern walkway?", the answer is already in the cloud — no return visit required.
• Safe stand-off. The plant could be scanned largely from accessible ground and platform positions, with the scanner reaching elevated, congested geometry a surveyor would otherwise have to climb to or enter under permit. On a live operation, removing people from those positions is a direct safety gain.
• One model, many uses. A single georeferenced capture serves the immediate expansion design and keeps paying off — clash detection, foundation design, shutdown planning and as-built verification all draw on the same dataset.

The judgement call was technique selection across the site. The pit shell and open infrastructure corridors are captured fastest and safest from the air; the dense, vertical, detail-critical plant demands the millimetre fidelity of a tripod-mounted terrestrial scanner. The programme used both, then merged them into one cloud — which only works if both datasets share the same control.

Equipment and methodology

Ground control and datum

Everything was referenced to GDA2020, MGA2020 Zone 51, with heights on AHD, matching the operation's mine grid so the resulting model dropped straight into the client's design environment. ISS established a network of survey control around the plant and across the broader site, observed by RTK off the mine's primary control with redundant occupations on each mark. Coded scan targets and checkerboard references were distributed through the plant to constrain the terrestrial registration, and ground control marks were laid out for the UAV capture. A set of independent check points was held back and not used in the adjustment, giving a blind measure of the final model's accuracy.

Control discipline is what lets two sensors become one model. ISS applies the rule that control should be two to three times more accurate than the survey it constrains; here the network resolved comfortably inside the centimetre, against a 25 mm model target.

Aerial capture

The UAV LiDAR captured the open pit, the ROM pad and the run-of-mine corridors — terrain and large structures where aerial point density is more than adequate and where flying is far faster and safer than walking. All flights were conducted under CASA Part 101 within ISS's Remotely Piloted Aircraft Operator's Certificate (ReOC) by a licensed Remote Pilot, inside the mine's site-specific RPAS procedures with a spotter coordinating over the radio to keep clear of haul trucks and the live working area.

Terrestrial scanning

The process plant was captured with a Leica RTC360 across a planned network of scan positions, each tied into the coded targets and the control network. The RTC360 was chosen for its speed and its on-board pre-registration, which let the field crew verify scan-to-scan link quality on site before demobilising — the brownfield equivalent of checking your work before you leave, because a missed scan position on a plant you cannot easily re-enter is a costly return trip. Setups were planned to see behind and around congested plant from multiple angles, minimising the occlusion shadows that plague single-direction capture.

Registration and modelling

Terrestrial scans were registered in Leica Cyclone REGISTER 360, then the registered cloud and the UAV LiDAR were brought together and constrained to the common control in a single georeferenced project. Registration of the plant held to 25 mm, verified against the withheld check points. From the unified cloud, ISS extracted the deliverables the design team needed: a coloured point cloud for navigation and visual reference, a 3D as-built model of the plant in the client's CAD environment, and surface models of the pit and infrastructure. Critical clearances and tie-in geometry were measured directly off the cloud and supplied as a marked-up reference set.

The result

The design team got what 12-year-old drawings could never give them: a measured, current, navigable model of the actual plant, tied to the coordinate system they were designing in. The expansion was routed against reality. New pipework was threaded through the existing congestion with verified clearances; the thickener was sited and its foundation set out against the as-built ground and structures; and crane access for the install was confirmed by checking swing paths and overhead clearances directly in the model — all before a single drawing went to the field.

The most valuable outcome was not the headline accuracy figure — it was the clashes that never happened. Several pipe routes and a structural connection in the original concept were found to interfere with existing plant once tested against the model. Catching those in CAD, weeks before fabrication, turned what would have been shutdown rework into a few hours of redesign. On a CIL plant where downtime carries the full cost of deferred gold production, that is the kind of saving that pays for the survey many times over.

Outcome and reuse of the model

The single LiDAR capture became a standing digital asset for the operation. After the expansion design, the same georeferenced cloud was reused to run formal clash detection on the detailed model, to design the thickener foundation against measured ground, and to support a tailings spigot survey — three further jobs that drew on the original data with no additional field mobilisation. This is the quiet economics of laser scanning a brownfield site: the marginal cost of each additional use of an existing, well-controlled model is close to zero.

There were safety and schedule dividends alongside the engineering ones. Capturing the plant as a point cloud meant the survey crew worked from safe, accessible positions rather than climbing into elevated and congested live process areas with instruments and tapes, cutting both working-at-height exposure and the isolation burden on operations. And because the model is current and traceable — datum, control residuals, registration accuracy and check results all documented — it gives the site a defensible baseline for the next shutdown, the next modification and the next as-built question, instead of starting from a drawing nobody trusts.

Frequently asked questions

How accurate is a LiDAR 3D model of a mine site?

It depends on the technique and what is being measured. Terrestrial laser scanning of a process plant, properly controlled and registered, holds to around 25 mm — fine enough to design pipe tie-ins and structural steel against. UAV LiDAR over a pit or infrastructure corridor is typically in the few-centimetre range vertically, which is appropriate for terrain and large structures. In this case study the plant model was registered to 25 mm against withheld check points, all referenced to GDA2020 / MGA2020 on AHD.

Why combine UAV LiDAR with terrestrial scanning instead of using one?

They solve different problems. UAV LiDAR captures large open areas — pits, ROM pads, corridors — quickly and safely from the air, but lacks the fidelity to model dense vertical plant. Terrestrial scanning delivers the millimetre detail a congested process plant needs but is slow over broad terrain. Using both, then merging them onto common control, gives one continuous model that is fast where speed matters and precise where precision matters.

Why not just use the existing as-built drawings?

On a long-running brownfield site the drawings are almost never current. Pipe runs get re-routed, platforms extended and equipment swapped during shutdowns, and the drawing set rarely keeps pace. Designing an expansion against outdated drawings is how clashes end up being discovered in the field during construction — the most expensive place to find them. A LiDAR model captures the plant as it actually is today.

What equipment and standards did ISS use?

A DJI Matrice 350 RTK with a Zenmuse L2 for the aerial LiDAR, flown under CASA Part 101 within ISS's ReOC by a licensed Remote Pilot, and a Leica RTC360 for terrestrial scanning of the plant. Scans were registered in Leica Cyclone REGISTER 360 and the combined cloud constrained to a control network in GDA2020 / MGA2020 Zone 51 on AHD, verified against independent check points.

Can the same approach work at other WA or interstate mine sites?

Yes. Any brownfield expansion, retrofit or as-built verification — across the Goldfields, the Pilbara or interstate operations — suits the same combined UAV and terrestrial LiDAR workflow. The method scales from a single process area to a full site, and the resulting model typically earns its keep several times over as later engineering work draws on the same controlled capture.

Talk to us about modelling your mine site

If you are planning a plant expansion, a retrofit or a shutdown and your as-built drawings can no longer be trusted, a controlled LiDAR 3D model removes the guesswork — and the clash risk that comes with it. Industrial Spatial Solutions delivers combined UAV and terrestrial laser scanning across Western Australian mine sites and interstate, referenced to GDA2020 / MGA2020 and AHD, registered to documented accuracy and delivered as a reusable digital asset. Call us on 0407 057 015 to scope your site and arrange a fixed-price quotation.

Industrial Spatial Solutions — site captured, plant modelled, design de-risked.

Related reading: 3D laser scanning services, UAV and aerial surveys, how accurate is 3D laser scanning, mining survey services.