What is Georeferencing?

TL;DR

Georeferencing is the process of tying spatial data — a point cloud, drone map, scan, or drawing — to a real-world coordinate system so every measured point sits in its correct geographic position. In Australia this means anchoring data to GDA2020, projecting it into MGA2020, and referencing height to the Australian Height Datum (AHD). Done well, georeferencing delivers absolute positioning to within 10-30mm; done poorly, a perfectly precise survey lands in the wrong place.

Key takeaways

• A georeferencing survey ties raw spatial data (point clouds, orthophotos, scans) to a national coordinate framework — in Australia, GDA2020 horizontal, MGA2020 grid, and AHD heights.
• Georeferencing controls absolute accuracy (where the data sits on Earth); registration controls relative accuracy (how scans fit together). They are different problems and both matter.
• Ground control points (GCPs) surveyed with GNSS or a total station are the link between local survey data and the national datum. A drone survey without GCPs can be metres out in absolute terms.
• Australia moved from GDA94 to GDA2020 in 2020 — a shift of roughly 1.8m north-east. Mixing the two datums on one site is a common and expensive error.
• Typical georeferenced absolute accuracy: 10-30mm for ground-controlled drone photogrammetry, 5-15mm for total-station-controlled laser scanning.

What is a georeferencing survey?

A georeferencing survey establishes the mathematical link between data captured in the field and a known, published coordinate system. Every survey instrument — a Leica RTC360 scanner, a DJI Matrice drone, a Trimble total station — initially records measurements in its own local frame. Georeferencing transforms those local measurements so that a pipe flange, pit wall, or building corner reports the same coordinate any other surveyor in the country would measure for that exact point.

The principle is straightforward: you measure a set of physical features whose coordinates are already known in the national datum, then solve the transformation (translation, rotation, and scale) that best maps your data onto those known points. Those known features are ground control points — surveyed targets, painted marks, or permanent survey marks (PSMs) maintained by state authorities.

Without georeferencing, a survey is internally consistent but geographically adrift. A drone orthophoto might be millimetre-sharp internally yet positioned two metres from reality — useless for stockpile reconciliation across surveys, clash detection against engineering design, or any work where your data must agree with someone else's.

Georeferencing vs registration: a critical distinction

These two terms are routinely confused, and the confusion causes real errors on site.

A laser scanning project can be perfectly registered — all 40 scan positions fitting together flawlessly — yet completely un-georeferenced, floating in an arbitrary local grid. Conversely, a poorly registered project that has been georeferenced will place the centre of mass in the right spot but carry internal distortion. Production-grade deliverables require both: tight registration and correct georeferencing against the national framework.

How georeferencing works

The workflow is consistent across drone, scanning, and conventional survey, though the tools differ.

The georeferencing process

1. Establish control. A licensed surveyor connects to nearby permanent survey marks or runs a GNSS network (static or RTK) to bring GDA2020 coordinates onto site. CORSnet-NSW, GPSnet (VIC), and AusCORS provide correction networks for centimetre RTK across most of Australia.
2. Place ground control points. Targets are distributed across and around the survey area — for drone work, a minimum of 5 GCPs, ideally one near each corner and one central, with more on large or tall sites. Each GCP is surveyed to GDA2020/MGA2020/AHD.
3. Capture the data. The drone flies its grid, the scanner occupies its stations, or the total station picks up detail — all referencing the same control.
4. Solve the transformation. Processing software (Agisoft Metashape, Trimble Business Center, Leica Cyclone REGISTER 360) computes the best-fit transformation onto the control. Software reports residuals at each GCP — the difference between the surveyed and the computed position.
5. Validate with check points. Independent points not used in the solution are compared against their known coordinates. This is the honest test of absolute accuracy.

Key point: The accuracy you can claim is set by your worst check point, not your best control point. A solution that hits 8mm RMS on control but 40mm on an independent check point is a 40mm survey. Surveyors who quote control residuals as if they were accuracy are misrepresenting the result.

Coordinate systems used in Australia

Georeferencing in Australia means working with three linked but distinct components.

GDA2020 (Geocentric Datum of Australia 2020) is the current national horizontal datum. It corrects the roughly 1.8m of tectonic drift that had accumulated since GDA94 was fixed — the Australian plate moves about 7cm per year north-east. MGA2020 (Map Grid of Australia 2020) is the projected grid derived from GDA2020; it is what appears on plans and what machine control systems consume. The Australian Height Datum (AHD) provides vertical reference, though many drone and GNSS workflows capture ellipsoidal height and apply the AUSGeoid2020 model to convert to AHD.

The single most common georeferencing error on Australian sites is mixing GDA94 and GDA2020 data — an as-built captured in GDA2020 overlaid on a legacy GDA94 design will sit a confusing ~1.8m out, often blamed on the instrument rather than the datum.

Where georeferencing matters

Mining and resources

Stockpile volumes on a Pilbara iron ore operation must be comparable month to month. If each drone survey is georeferenced to the same GDA2020/MGA2020 control, volumes reconcile; if not, apparent tonnage swings reflect positioning drift, not real material movement. Pit design, haul road grading, and tailings dam monitoring all depend on every survey landing on the same grid.

Construction and civil

Setout and as-built verification require survey data to agree with the engineer's design model. Machine control on a dozer or grader reads MGA2020 coordinates directly — georeferencing is what lets the blade cut to design. A formwork or pile-position survey that is not correctly georeferenced cannot be compared against the design tolerance.

Industrial and shutdowns

Laser scans of a processing plant or refinery, captured across a shutdown, must georeference to plant grid so that new spool fabrication, clash detection, and tie-in design align with both existing steel and the design intent. Scan data floating in a local frame cannot be merged with the client's plant coordinate system.

Asset and infrastructure monitoring

Deformation monitoring of a wharf, dam, or crane rail relies on repeat surveys georeferenced to the same stable control network. Movement is only meaningful if the reference frame itself has not moved between epochs.

Georeferencing methods compared

In practice, the methods combine. A drone survey commonly uses PPK GNSS for the bulk solution with GCPs as the absolute anchor and independent check points for validation. A scanning project ties scan targets to a total station traverse that is itself connected to PSMs or a GNSS-observed GDA2020 base.

Frequently asked questions

What is georeferencing in surveying?

Georeferencing in surveying is the process of assigning real-world coordinates to spatial data so that every point sits in its correct geographic position. In Australia, this means tying the data to GDA2020 horizontally, MGA2020 as the working grid, and AHD for height. It is what makes a survey comparable with other surveys, engineering designs, and machine control systems.

What is the difference between georeferencing and registration?

Georeferencing controls absolute accuracy — where the whole dataset sits on Earth, relative to a national datum. Registration controls relative accuracy — how individual scans or images align to one another. A project can be well registered but un-georeferenced (internally perfect, geographically wrong), so production deliverables need both.

How accurate is a georeferenced survey?

Absolute accuracy depends on method. Ground-controlled drone photogrammetry typically achieves 10-30mm, RTK/PPK drone work 20-50mm, and total-station-controlled laser scanning 5-15mm. The honest measure of accuracy is the residual at independent check points, not the control points used in the solution.

What coordinate system is used for georeferencing in Australia?

Australia uses GDA2020 as the national horizontal datum, MGA2020 (in UTM zones 49 to 56) as the projected working grid, and the Australian Height Datum (AHD) for heights. GDA2020 replaced GDA94 in 2020, correcting roughly 1.8m of accumulated tectonic drift, so data must not be mixed between the two datums.

Why do drone surveys need ground control points?

A drone's onboard GNSS alone can leave a map metres out in absolute terms. Ground control points — targets surveyed to GDA2020/MGA2020/AHD — anchor the imagery to the national framework and let the processing software solve and validate the transformation. Without GCPs (or precise PPK with check points), a drone survey cannot be relied upon for absolute positioning.

What to do next

Georeferencing is the difference between a survey that merely looks precise and one that is actually correct on the national grid. If your project involves repeat measurement, comparison against design, machine control, or merging data with a client's coordinate system, georeferencing to GDA2020 / MGA2020 / AHD is non-negotiable.

1. Confirm your required datum and grid — most Australian projects specify GDA2020, MGA2020, and AHD, but check the client's plant or design framework first.
2. Specify absolute accuracy and check-point validation — ask for residuals at independent check points, not just control.

Call ISS on 0407 057 015 for a fixed-price quote. We georeference drone, laser scanning, and conventional survey data to GDA2020, MGA2020, and AHD across Australia, with independent check-point validation on every deliverable.