Ground Control Points (GCPs) remain one of the most important tools in aerial photogrammetry. Even with the rise of RTK and PPK drones that capture accurate geotags in real time, GCPs provide the reliable control that ensures an orthophoto or 3D model truly matches the ground. By surveying and marking precise points on site, we can “lock” the imagery to a known coordinate system, guaranteeing consistent accuracy across a project.
But good photogrammetry is not just about GCPs. It’s equally important to check the quality of the outputs, and that’s where Quality Control Points (QCPs) come in. These are points on the ground that have been surveyed like GCPs but are not used in processing the imagery. Instead, they serve as independent checkpoints, verifying the positional accuracy of the final orthophoto or 3D model. If the QCPs line up closely with their real-world coordinates, it’s a strong indication that the dataset has been processed correctly. If they don’t, it’s a warning flag that something has gone wrong — perhaps poor target marking, insufficient overlap, or a weak terrain model.
This is why many surveyors still rely on a blend of GCPs and QCPs. GCPs provide the structural framework that anchors the model, while QCPs confirm its accuracy after processing. Together they give confidence that the outputs can be trusted for engineering, design, or construction decisions.
It’s also worth remembering that resolution and accuracy are not the same thing. Just because an orthophoto has a fine pixel resolution — say 2 cm/pixel — does not guarantee that every feature is positioned within 2 cm on the ground. True positional accuracy depends on strong control, well-distributed GCPs, and an appropriate digital elevation or terrain model (DEM/DTM). Without these, even the sharpest imagery can distort distances, slopes, or alignments. In other words: resolution gives you detail, but control and elevation data give you trust.
Key Points from This Article
- GCPs anchor orthophotos to the real world by fixing latitude, longitude, and height, transformed into Easting, Northing, and Elevation.
- QCPs provide independent checks, confirming that processed models align with surveyed ground truth.
- Balanced distribution is critical — place GCPs around edges, across the centre, and at elevation highs and lows, with ~100 m separation.
- Target size matters — use checkerboards or painted crosses at least 5–10× the GSD for reliable marking.
- Clustering weakens accuracy — spread GCPs evenly to prevent distortions in the model.
- RTK/PPK reduces but does not replace control — hybrid workflows with GCPs and QCPs are most efficient.
- Case study at Chatterley Valley showed how 20+ GCPs and independent QCPs secured sub-centimetre accuracy for weekly cut-and-fill calculations.
- Analogy to remember: GCPs are tent pegs holding the orthophoto in shape, QCPs are the checks that confirm it’s pitched correctly.
What GCPs and QCPs Do in Your Workflow
Ground Control Points (GCPs) are surveyed reference points on the ground with known coordinates. They anchor the aerial survey to the real world by fixing its absolute position in latitude, longitude, and height. Since drones inherently measure in geographic coordinates (Lat/Long/Height), these values must be transformed into the working coordinate system of the project — typically X, Y, Z or Easting, Northing, Elevation. By doing this, GCPs reduce GPS drift from metres down to centimetres, constrain the geometry of the photogrammetric block, and prevent distortions such as tilt, bowing, or misalignment.
Quality Control Points (QCPs) serve a different but equally vital role. They are surveyed in the same way as GCPs but are deliberately excluded from the processing workflow. Instead, they act as independent validation markers. Once the orthophoto or 3D model has been processed, the measured positions of the QCPs can be compared with their surveyed coordinates. If the differences are small and within tolerance, it confirms that the processing has been done correctly. If they are larger than expected, it’s a warning that control was insufficient, marking was poor, or that more refinement is needed. QCPs provide the evidence that your processed model is as accurate as you claim it to be.
Together, GCPs and QCPs deliver not just accuracy, but proof of accuracy. This is particularly valuable when handing over deliverables to clients, contractors, or regulators, because it demonstrates both control during capture and quality assurance afterwards.
An easy way to visualise this is to imagine an orthophoto as the canvas of a tent. The GCPs act like the tent pegs and poles, holding the canvas in the correct position and shape. The QCPs are the inspection points, where you check that the canvas is stretched evenly and firmly. Without the pegs and poles, the tent would sag and distort; without the inspection points, you wouldn’t know whether it had been pitched correctly.
How Many GCPs Do You Need?
There isn’t a one-size-fits-all answer to the question of how many Ground Control Points (GCPs) are required for an aerial survey. The right number depends on the size of the site, the level of accuracy required, and the topography of the ground. However, there are some clear principles that make planning easier.
A good rule is to surround the survey area with GCPs placed around the edges and then dot additional points across the internal area. This ensures that the photogrammetric model is pinned down both at the boundaries and throughout the interior. If you only place points on the edges, the centre can deform like a loose sheet; if you only place points in the middle, the edges can drift. A balanced spread is essential.
Topography plays a critical role. On flat ground, fewer points may be sufficient because the geometry is simple, and the model is less likely to warp. On sites with significant elevation changes — such as embankments, stockpiles, or valleys — you need more GCPs placed at the highs and lows. This stabilises the vertical accuracy of the orthophoto and prevents the model from “sagging” or “tilting” in those areas.
One way to visualise this is to think of the orthophoto like the fabric of a tent stretched over a frame. Each GCP is a peg or pin holding the canvas in place. If you only pin down one side, the fabric will flap and distort. But if you space the pegs evenly around the edges and add a few in the middle, the whole tent remains taut and true.
Clustering GCPs in one area is counterproductive. Too many close together add little value, while leaving other areas uncontrolled. Instead, aim for at least 100 metres of separation between GCPs on most sites. On larger or more complex areas, you may need tighter spacing, but the principle of even distribution always applies.
As a practical guide:
- Small, flat sites (under 5 hectares) may be controlled with 5–8 well-placed GCPs.
- Medium sites (5–20 hectares) often require 8–15 points, with extras added for topographic features.
- Large or complex sites (20 hectares or more, or with significant height differences) may need 15–30 or more, depending on the required accuracy.
The key is not just the number, but the coverage. If you can imagine drawing a web that connects all your GCPs without leaving large holes, your layout is likely to give good results.
Rule of Thumb for GCP Placement
- Place around the edges of the site first.
- Add extra points across the middle for stability.
- Include high and low elevations if the terrain varies.
- Keep ~100 m separation between points where possible.
- Avoid clustering – spread them evenly.
- Always add a few extra QCPs (checkpoints) for quality assurance.
What Size Should Targets Be?
The effectiveness of a Ground Control Point (GCP) depends not just on its position, but also on how easily it can be seen and marked in the imagery. Targets should be large, high-contrast, and unambiguous in shape so that their centres can be identified accurately in multiple overlapping photos.
The most common choice is the checkerboard-style GCP, usually black and white squares in a cross pattern. These work well because the intersection of the pattern creates a clear, precise point that is easy to spot, even at higher altitudes. Checkerboards are durable, reusable, and often mounted on boards, tarpaulins, or rigid panels so they can be placed and recovered easily.
However, a GCP doesn’t always need to be a manufactured target. In practice, any distinguishable feature that can be surveyed can serve the same purpose, provided it is visible in enough images. On some projects, permanent checkerboards aren’t practical because they can be damaged by plant, covered in mud, or destroyed once work begins. In these situations, a painted cross on hardstanding, tarmac, or concrete can work just as well. These painted crosses are typically large — at least 600 mm long in each direction — to ensure they are clearly visible in imagery. The bright colour against the background provides a strong centre point to mark during processing.
Relating Target Size to Image Resolution
When deciding on target size, always relate it to the Ground Sampling Distance (GSD) of your survey. A good rule of thumb is that a target should be 5–10 times the GSD. For example:
- At a GSD of 2 cm/pixel, a target should be 10–20 cm wide.
- At a GSD of 5 cm/pixel, a target should be at least 25–50 cm wide.
This ensures that the centre of the target spans enough pixels to be marked reliably in the software. Targets that are too small relative to the GSD will appear blurred, making their centres difficult to pinpoint and reducing accuracy.
Other Considerations
- Contrast: Use colours that stand out against the background (white on asphalt, black on light surfaces, or fluorescent paint if necessary).
- Durability: For long-term projects, portable boards or tarpaulins can be moved and reused. For shorter-term or temporary control, paint is often faster and more practical.
- Visibility in multiple frames: Ensure each GCP is captured in at least 8–12 images from different angles to maximise processing accuracy.
In short: while checkerboard targets are the industry standard, don’t hesitate to adapt to site conditions. Whether it’s a painted cross, a tarpaulin, or even a natural but distinct feature, the goal is the same — to provide a clear, measurable point that ties your aerial survey to the ground with confidence.
Where to Place GCPs and QCPs
Placing Ground Control Points (GCPs) and Quality Control Points (QCPs) correctly is just as important as deciding how many you need. Good placement ensures that the photogrammetric model is well anchored across the entire site, while poor placement can leave areas of the model distorted or floating free.
Distribute Around the Perimeter
Start by placing GCPs around the edges of the survey area. Think of these as the tent pegs that pin down the outer edges of your orthophoto. Without perimeter points, the “fabric” of the orthophoto is free to stretch or drift at the boundaries, leading to poor alignment where accuracy often matters most.
Add Points Across the Interior
Once the perimeter is secured, dot GCPs through the internal areas. These act like poles pushing the canvas upwards and ensuring the surface is held firmly across the middle. Without interior points, the centre of the orthophoto can sag or warp, even if the edges are locked down.
Cover Elevation Extremes
If the site has significant height differences — embankments, cuttings, stockpiles, or deep excavations — place GCPs at the highest and lowest points. This stabilises vertical accuracy and prevents the model from “tilting” or distorting in areas of steep relief.
Avoid Clustering
Don’t cluster too many GCPs in one place. A tight group of points adds little value while leaving other areas uncontrolled. Spread them evenly across the site, aiming for at least 100 metres of separation where possible. This helps create a balanced, well-tied network that anchors the orthophoto across its full extent.
Place QCPs Independently
QCPs should be separate from the GCP network. Distribute them across the site in areas not already covered by GCPs — ideally near the centre and at elevation changes. Their job is to test whether the model holds its shape independently of the control points. By keeping them separate, you ensure your accuracy checks are truly unbiased.
Corridor Surveys
For long, narrow sites like roads, railways, or pipelines, place GCPs in a zig-zag pattern down the corridor. Add pairs of points at each end to lock the geometry in place and prevent drift along the length. QCPs can be slotted into the same zig-zag pattern but should be left out of the processing so they can serve as independent checks.
Visualising Placement
Again, picture your orthophoto as a tent canvas:
- The perimeter GCPs are the pegs holding the edges down.
- The interior GCPs are the poles keeping the fabric stretched evenly.
- The QCPs are your inspection points, where you check that the canvas hasn’t sagged, twisted, or pulled loose once the tent is up.
With this layout, your orthophoto will be both structurally sound and verifiably accurate.
GCPs, QCPs, and RTK/PPK: A Balanced Approach
Modern drones equipped with RTK (Real-Time Kinematic) or PPK (Post-Processed Kinematic) positioning systems have dramatically improved aerial surveying. By correcting the GPS signal in real time or after flight, these drones can achieve positional accuracy down to a few centimetres — often without the need for a dense network of Ground Control Points (GCPs).
But while RTK and PPK reduce the reliance on GCPs, they do not remove the need for control and quality assurance. Relying solely on onboard positioning is a bit like trusting the pegs of your tent will stay in place without checking if the ropes are taut — it might look fine at first glance, but without verification, you can’t be sure it’s holding properly.
Why You Still Need GCPs
Even with RTK or PPK drones, a small number of strategically placed GCPs helps “lock” the orthophoto or 3D model into the local grid system. They act as anchor points that confirm the drone’s onboard corrections are translating correctly into the chosen coordinate system (Easting/Northing/Elevation). On complex sites or where local accuracy is critical — for example, when tying into engineering drawings — these GCPs provide peace of mind.
The Role of QCPs in RTK/PPK Surveys
This is where Quality Control Points (QCPs) become invaluable. By leaving some surveyed points out of the processing and using them purely as checkpoints, you gain an independent measure of accuracy. If your RTK- or PPK-driven dataset holds up against QCPs within a few centimetres, you can confidently demonstrate that the survey meets specification. If not, you’ll know the system drifted, and you’ll have the evidence to reprocess or re-fly before costly mistakes are made.
Finding the Right Balance
The most efficient workflow is often a hybrid approach:
- Use RTK or PPK positioning to reduce the total number of GCPs needed.
- Place a handful of well-distributed GCPs to tie the dataset securely to the ground.
- Always include QCPs as independent checks, so you can prove the accuracy of your outputs to clients, contractors, or regulators.
This combination gives you the speed and efficiency of RTK/PPK while still providing the security and accountability of traditional control.
Analogy to Remember
Think of your survey like pitching a tent on uneven ground:
- RTK/PPK gives you pre-measured guy ropes, helping you set up quickly with a good initial fit.
- GCPs are the strong pegs you hammer in at key points to make sure the tent is anchored to the ground.
- QCPs are the inspection checks afterwards — walking around the tent, tugging on the ropes, and making sure everything is secure.
Use all three, and your tent — like your orthophoto — won’t just look right, it will stand firm under scrutiny.
Case Study: Weekly Drone Surveys at Chatterley Valley
Between June and December 2024, the regeneration project at Chatterley Valley provided the perfect example of how GCPs and QCPs add value to aerial surveying. Over this six-month period, 26 weekly drone surveys were completed using a DJI Mavic 3 Enterprise equipped with an RTK module.
Each flight covered approximately 446,000 m² (around 1.44 km long by 400 m wide), producing more than 2,000 images per survey at a ground sampling distance (GSD) of ~1.5 cm/pixel with 75% side and 80% forward overlap. On paper, the RTK corrections alone should have delivered a very precise dataset — but for a project of this scale and importance, relying on drone GNSS alone wasn’t enough.
How GCPs Were Used
More than 20 GCPs were placed around the site perimeter and through the internal working areas. These were carefully positioned to tie the survey tightly to the local grid system and to account for the site’s varied topography. Points were included at the highest ground levels, near embankments, and at the lowest excavations. This distribution stabilised the geometry of the orthophotos and reduced the risk of “warping” in the vertical dimension.
The Importance of QCPs
Alongside the GCPs, several Quality Control Points (QCPs) were also surveyed but deliberately left out of the processing. Once each weekly orthophoto was generated, these QCPs were checked against their surveyed coordinates. The differences consistently came back within sub-centimetre tolerances, giving independent proof that the outputs were not just sharp, but accurate. If any QCP had shown unexpected errors, it would have been an early warning that something was wrong — prompting a re-check before contractors acted on the data.
Why It Mattered
The Chatterley Valley project required precise cut-and-fill calculations to monitor earthworks over time. Even a few centimetres of drift from week to week could have introduced serious errors into the volume comparisons. By combining RTK positioning with a strong GCP framework and validating everything against QCPs, the survey team had confidence that each dataset was directly comparable.
For the client, John F Hunt Regeneration Ltd, this meant every week’s data could be trusted to make decisions about material movements, programme scheduling, and subcontractor performance. The use of GCPs and QCPs ensured the survey outputs were not only detailed but also defensible — the kind of data that could withstand technical or contractual scrutiny.
Step-by-Step Checklist (Field → Office)
Before you fly
- Define accuracy requirement and choose flight altitude / GSD.
- Decide control strategy (RTK/PPK + lean GCPs + QCPs, or GCP-heavy if GNSS is poor).
- Prepare targets sized 5–10 × GSD (checkerboards, boards, or painted crosses).
In the field
- Place GCPs evenly: around edges, across centre, and at elevation highs/lows.
- Place QCPs separately: keep them out of processing for QA checks.
- Aim for >100 m spacing between points; avoid clustering.
- Survey all targets (Lat, Long, Height) and record as Easting, Northing, Elevation.
- Photograph and document each point (ID, description, photo).
During processing
- Use GCPs for alignment and block adjustment.
- Exclude QCPs — use them only to validate the model afterwards.
- Check QCP residuals (horizontal + vertical); report RMS errors in deliverables.
- Export outputs (orthophotos, DEMs, point clouds) with accuracy notes included.
Common Mistakes to Avoid
Even with modern drones and software, mistakes in planning or executing your control network can undermine the accuracy of your entire survey. Here are some of the most common pitfalls — and why they matter:
Clustering GCPs in One Area
Placing too many GCPs close together feels like extra security, but it adds little value. The software already knows how to align images locally — what it needs are anchors spread across the site. If all your GCPs are in one corner, the rest of the orthophoto can drift, tilt, or distort. Spread them evenly across the footprint to tie down the whole area.
Forgetting QCPs
A model can look perfectly aligned with its GCPs, but without independent Quality Control Points, you have no way to verify it. Skipping QCPs removes your only independent check that the orthophoto or DEM is accurate. Always keep a few surveyed points out of processing and use them purely as checkpoints.
Targets Too Small for the GSD
If a GCP target is only a few pixels wide in your imagery, you won’t be able to mark its centre accurately. This introduces noise and inconsistency into the adjustment. Make sure every target is 5–10 times larger than the GSD so it can be confidently identified in multiple images.
Control Only at the Edges
It’s tempting to just ring the perimeter with GCPs, but without interior points, the middle of the model can sag like an unsupported tent canvas. Always add a few GCPs through the centre to stabilise the internal geometry.
Ignoring Elevation Changes
On sites with big height differences — embankments, stockpiles, valleys — placing GCPs only on flat ground will leave your vertical accuracy weak. Include GCPs at both the highest and lowest elevations to prevent tilt or warping.
No QA Reporting
Even if you collect QCPs, forgetting to report their residuals leaves clients in the dark about how accurate the deliverables really are. Always provide the RMS error values for your QCPs in both horizontal and vertical directions. This gives your client confidence — and protects you if your work is ever audited.
Summary
Accurate aerial surveys depend on more than just high-resolution imagery. While RTK and PPK drones have raised the baseline for positional accuracy, Ground Control Points (GCPs) and Quality Control Points (QCPs) remain essential tools for producing datasets that are both precise and verifiable.
GCPs act as the anchors of your orthophoto, locking it to the real-world coordinate system in latitude, longitude, and height, and then transforming those into the working grid of Easting, Northing, and Elevation. QCPs provide the proof that the outputs hold true by checking against surveyed points that were not used in processing. Together, they ensure that the final product isn’t just sharp, but trustworthy.
The lesson from projects like Chatterley Valley is clear: a successful workflow uses a balance of technologies. RTK positioning provides efficiency, GCPs provide structure, and QCPs provide confidence. Think of it as pitching a tent — RTK gives you the ropes, GCPs are the pegs holding it down, and QCPs are the checks you make afterwards to be sure it’s pitched correctly.
By investing the time to plan and execute a robust control strategy, you deliver data that engineers, contractors, and clients can act on with certainty. In aerial surveying, resolution shows the detail — but control and quality assurance deliver the reliability that truly matters.