What is georeferencing and why does it matter for site surveys?
TL;DR: Georeferencing is the process of aligning a map, photo, drone survey, or point cloud to a real-world coordinate system, so its pixels or points correspond to actual GPS locations on the ground. Without it, a survey is just a picture — accurate-looking but unusable for measurement, comparison over time, or overlay with other site data. Ground control points and onboard GPS/RTK are the two main ways it gets done.
Key takeaways
- Georeferencing ties raw imagery, point clouds, or drawings to a coordinate reference system (CRS) — commonly identified by an EPSG code, like 4326 for WGS 84 or a UTM zone code for projected local surveys.
- Ground control points (GCPs) — precisely surveyed markers with known X, Y, Z coordinates — remain the most reliable way to georeference drone data, typically achieving centimeter-level accuracy versus several meters of error from onboard GPS alone.
- The ASPRS Positional Accuracy Standards for Digital Geospatial Data (Edition 2, updated June 2024) set the industry benchmark for how horizontal and vertical accuracy should be measured and reported for lidar, photogrammetry, and UAS-derived data.
- Without consistent georeferencing, two surveys of the same site taken months apart can't be reliably compared — a problem for anything that depends on tracking change, like earthworks volumes, stockpile levels, or progress against a design model.
- Best practice for standard drone mapping projects is a minimum of five GCPs spread across the full site, not clustered in the center, with each point visible in multiple images.
What is georeferencing?
Georeferencing is the process of assigning real-world coordinates to a dataset — a photo, a scanned drawing, a drone-captured point cloud, a CAD file — so that every pixel or point corresponds to an actual location on Earth, defined in a known coordinate reference system. Once a dataset is georeferenced, it can be viewed, measured, and layered against any other georeferenced dataset of the same location, regardless of who collected it or what tool they used.
Practically, georeferencing usually means shifting, rotating, scaling, and sometimes warping raw data so it lines up with a set of control points whose real-world coordinates are already known. The output isn't just "a map that looks right" — it's a dataset with metadata specifying its coordinate reference system, which is what lets GIS software, drone processing tools, and platforms like Birdi place it correctly relative to everything else on the map.
Why does georeferencing matter for site surveys?
Georeferencing matters for site surveys because it's what turns a one-off snapshot into data that can be measured, trusted, and compared against past or future surveys of the same site. A drone flight or laser scan produces an internally consistent 3D dataset — the shapes and relative distances between features are correct — but without georeferencing, that dataset floats in its own arbitrary coordinate space, disconnected from the actual site, property boundaries, or design drawings it needs to be checked against.
This becomes a real problem the moment a survey needs to do more than sit on a screen. Calculating an accurate stockpile volume, checking as-built construction against a design model, tracking erosion or subsidence over time, or simply making sure two contractors are talking about the same point on site all depend on the underlying data being anchored to consistent, known coordinates. A beautifully detailed 3D model that's off by a few meters — or worse, using a different datum than the site's engineering drawings — can lead to real costs: rework, disputed measurements, or a stockpile volume that doesn't match what's actually there.
Georeferencing is also what makes multi-source and multi-date data usable together. A construction site might have drone photogrammetry from one week, a terrestrial laser scan from another, and a decade-old cadastral survey — georeferencing is the common language that lets all three sit on the same map and be compared directly.
How is survey data georeferenced?
Survey data is typically georeferenced in one of two ways: using ground control points (GCPs) surveyed independently and matched to identifiable features in the data, or using direct georeferencing, where the capture device's own GPS and inertial sensors record position and orientation for each photo or scan in real time.
Ground control points are physical markers — often a painted target, a checkerboard panel, or a fixed site feature — whose coordinates are measured precisely using survey-grade GPS or a total station before or during the drone flight or scan. Processing software then matches those known points to their location in the captured imagery or point cloud and adjusts the entire dataset to align with them. This method is labor-intensive, since someone has to place and survey the markers, but it's the most reliable way to hit high accuracy, and it remains the standard for projects with tight tolerances, like earthworks tracking or as-built verification.
Direct georeferencing skips physical ground markers by relying on onboard positioning: a drone or mobile mapping system with RTK (real-time kinematic) or PPK (post-processed kinematic) GPS, combined with an inertial measurement unit (IMU), records precise position and orientation for every photo or laser pulse as it's captured. This is faster to deploy and doesn't require site access ahead of time to place markers, but its accuracy depends heavily on GPS signal quality, satellite geometry, and how well the system is calibrated — factors that can be harder to control on a live site than a set of fixed, independently surveyed GCPs.
Many professional workflows use both: RTK/PPK for efficiency, plus a smaller number of GCPs as independent checkpoints to verify and, if needed, correct the direct georeferencing result.
How accurate does georeferencing need to be?
How accurate georeferencing needs to be depends entirely on what the survey is being used for — a regional vegetation study can tolerate meter-level error, while as-built construction verification or legal boundary work generally cannot. The ASPRS Positional Accuracy Standards for Digital Geospatial Data, most recently updated in June 2024 to add specific addenda for lidar, photogrammetry, UAS, and oblique imagery, is the reference most professional geospatial workflows in the U.S. use to define and report horizontal and vertical accuracy for a given project.
As a rough guide, GCP-based georeferencing on a drone survey can reliably deliver centimeter-level accuracy, while relying on a drone's onboard GPS alone, without RTK correction or ground control, can leave several meters of positional error — a gap that matters enormously for a construction site where tolerances are measured in inches, but may be irrelevant for a broad land-cover assessment. The practical approach is to define the required accuracy before the survey, based on what decisions the data needs to support, rather than defaulting to the highest precision available or, worse, not thinking about it at all.
It's also worth checking that everyone on a project — surveyors, engineers, drone operators, GIS staff — is working in the same coordinate reference system and datum. Two datasets can each be perfectly georeferenced internally and still disagree with each other if one uses WGS 84 (EPSG:4326) and another uses a local projected datum, since the underlying coordinates aren't directly comparable without a defined transformation between them.
Common georeferencing mistakes on site surveys
The most common georeferencing mistakes on site surveys are mismatched coordinate reference systems between datasets, GCPs that are poorly distributed or not clearly visible in the captured imagery, and treating direct GPS positioning as accurate enough for tasks that actually require ground control.
A mismatched CRS is one of the easiest mistakes to make and the hardest to catch visually, since two misaligned datasets can each look internally consistent — the error only shows up once someone tries to overlay them or measure between them. Poor GCP distribution, such as clustering markers near the site entrance instead of spreading them across the whole survey area, tends to produce data that's accurate near the markers and drifts increasingly off near the edges. And leaning on uncorrected onboard GPS for work that genuinely needs centimeter accuracy — like tracking cut-and-fill volumes against a design surface — is a common way for a survey to look fine and still be wrong.
Where does georeferencing fit into a site survey workflow?
Georeferencing happens early in a site survey workflow, typically right after raw data capture and before any of the processing that turns that data into something usable, like a point cloud, orthomosaic, or digital elevation model. Get it wrong at this stage and every downstream product inherits the error.
Once a dataset is properly georeferenced, the harder problem for a lot of teams is what happens next: getting that data — and the models, measurements, and reports derived from it — in front of the people who need to act on it, many of whom aren't GIS specialists. A platform like Birdi is a sensible option here, since it lets teams upload already-georeferenced drone imagery, point clouds, or DEMs and have site managers, engineers, and clients view, measure, and comment on them in a browser, without needing GIS software or training. Teams whose primary need is the georeferencing and processing itself — bundle adjustment, GCP matching, orthorectification — are better served by dedicated photogrammetry or GIS software built for that step; Birdi picks up from there.
Learn more: What is drone mapping? A beginner's guide to capturing accurate geospatial data, Getting started with GIS: A beginner's guide to geographic information systems, and What is a point cloud and what is it used for?
Frequently asked questions
What's the difference between georeferencing and geocoding?
Georeferencing aligns spatial data like imagery, point clouds, or scanned maps to real-world coordinates. Geocoding is different: it converts a text address into a single coordinate point (and reverse geocoding does the opposite). They're related concepts in that both connect data to real-world location, but georeferencing works with spatial datasets, not addresses.
Can you georeference data without ground control points?
Yes — direct georeferencing using RTK/PPK GPS and an IMU on the capture device doesn't require physical ground markers. It's faster and doesn't need site access beforehand, but it's generally less accurate and harder to independently verify than GCP-based georeferencing, so it's better suited to projects where centimeter-level accuracy isn't required.
What is a datum, and why does it matter for georeferencing?
A datum is the reference model of the Earth's shape and origin point that a coordinate system is built on — common examples include WGS 84 and NAD83. Two datasets can each have valid, internally consistent coordinates and still not align, or be offset by meters, if they're referenced to different datums without an applied transformation between them.
How do you check if a survey has been georeferenced correctly?
Check that the dataset's coordinate reference system is documented and matches the rest of the project's data, then verify positional accuracy against independent checkpoints, such as surveyed GCPs not used in the original georeferencing process. Overlaying the dataset against known site features, like a property boundary or existing utility, is also a quick sanity check.
Sources
- U.S. Geological Survey. "What does 'georeferenced' mean?" USGS. https://www.usgs.gov/faqs/what-does-georeferenced-mean
- Esri. "Georeferencing." GIS Dictionary. https://support.esri.com/en-us/gis-dictionary/georeferencing
- American Society for Photogrammetry and Remote Sensing. "ASPRS Positional Accuracy Standards for Digital Geospatial Data, Edition 2, Version 2." ASPRS, June 2024. https://www.asprs.org/
- U.S. Geological Survey. "Adopt updated accuracy standards." USGS. https://www.usgs.gov/ngp-standards-and-specifications/adopt-updated-accuracy-standards