Drone technology has transformed the way Surveyors collect, process and analyse spatial data. With an unmanned aerial vehicle (UAV), a Surveyor can capture hundreds or thousands of overlapping images of a site and process them into orthomosaics, digital surface models (DSMs), digital terrain models (DTMs) and three-dimensional models.
However, an important question arises: How do we ensure that the maps and models produced from drone images are positioned accurately on the ground?
This is where Ground Control Points (GCPs) become important.
Although drone imagery can produce impressive visual results, the quality of a photogrammetric product depends not only on image resolution or the number of photographs captured, but also on how well the resulting data is referenced to real-world coordinates.
Understanding GCPs is therefore essential for Surveyors who want to produce reliable, measurable and defensible geospatial information.
1. What Are Ground Control Points (GCPs)?
Ground Control Points are clearly identifiable points on the Earth’s surface whose coordinates have been determined through surveying using an appropriate reference system and suitable measurement procedures.
In drone photogrammetry, these points are identified in the aerial images and used during photogrammetric processing to help align the reconstructed model with its correct real-world position, orientation and scale.
A typical GCP may consist of a high-contrast target placed on the ground, such as a black-and-white checkerboard or cross-shaped marker. The target is positioned so that it can be clearly seen in several overlapping drone images.
The Surveyor determines the coordinates of the target using suitable equipment and methods, such as:
* Global Navigation Satellite System (GNSS) receivers.
* Real-Time Kinematic (RTK) GNSS surveying.
* Post-Processed Kinematic (PPK) workflows, where appropriate.
* Total stations, particularly where satellite visibility is obstructed.
The measured coordinates are subsequently entered into photogrammetry software, where the corresponding points are marked in the captured images.
The fundamental principle is simple: the drone captures the images, while surveyed ground coordinates help connect the reconstructed model to the real world.
2. Why Are Ground Control Points Important?
Without adequate ground referencing, a photogrammetric model may look visually correct while containing positional errors, distortion or an incorrect relationship to the required coordinate system.
GCPs help address these issues in several ways.
A. Improving positional accuracy
GCPs help constrain the photogrammetric model to known ground coordinates. When properly distributed and measured, they can improve the absolute positional accuracy of the resulting map or model.
This is particularly important for cadastral support, engineering surveys, construction monitoring and infrastructure mapping, where the location of a feature matters as much as its appearance.
B. Supporting correct scale and orientation
Known ground coordinates help establish the model’s relationship to the real-world coordinate reference system. This supports reliable measurements of distances, areas and elevations, subject to the quality of the imagery, processing and control.
C. Reducing systematic errors
Photogrammetric processing may produce systematic distortions due to camera geometry, image network configuration, weak image overlap, terrain variation or inadequate positioning information.
Well-planned GCPs can help constrain these errors, although they cannot compensate for every weakness in flight planning, image quality or camera calibration.
D. Increasing confidence in survey deliverables
Clients often require spatial products that can be compared with existing survey records, engineering drawings or other geospatial datasets.
Appropriate ground control helps make these comparisons meaningful and improves confidence in the resulting deliverables.
3. How Do GCPs Work in Drone Photogrammetry?
GCPs form part of a wider photogrammetric workflow. They do not operate independently of image acquisition and processing.
A typical workflow involves the following stages:
Step 1: Plan the drone flight
Define the project boundary, flight height, image overlap, camera settings and required ground sampling distance (GSD). Select an appropriate coordinate reference system and plan the ground control survey.
Step 2: Establish and survey the GCPs
Place suitable targets at selected locations across the project area. Determine their coordinates using an appropriate surveying method and record the coordinate reference system, measurement method and relevant quality information.
Step 3: Capture overlapping aerial images
Fly the drone according to the approved mission plan. Ensure that the images provide sufficient overlap and suitable coverage of the control targets.
Step 4: Process the imagery
Import the photographs into photogrammetry software such as Agisoft Metashape or Pix4Dmapper. The software identifies common features across images and performs image alignment and three-dimensional reconstruction.
Step 5: Mark the GCPs in the images
Identify each surveyed target in the images where it appears. Enter its known coordinates and review the quality of the image measurements.
Step 6: Optimise the photogrammetric model
Use the GCP observations during bundle adjustment or the relevant optimisation procedure to refine the model’s relationship to the surveyed ground coordinates.
Step 7: Validate the results
Evaluate the processing statistics and, importantly, compare the outputs against independent Check Points that were not used to adjust the model.
Step 8: Generate the final deliverables
Produce the required orthomosaic, DSM, DTM, point cloud, contour map or three-dimensional model, depending on the project’s objectives.
4. Ground Control Points vs Check Points
One of the most important distinctions in photogrammetry is the difference between a Ground Control Point and a Check Point.
Although both have known ground coordinates, they serve different purposes.
Ground Control Points (GCPs) are incorporated into the adjustment process to help constrain the photogrammetric model.
Check Points are independent surveyed points withheld from the adjustment and used to evaluate the positional accuracy of the final output.
For example, a Surveyor may establish 12 surveyed targets across a site. Eight may be designated as GCPs, while four are reserved as Check Points.
The appropriate number and arrangement depend on the project size, terrain, image geometry, accuracy requirements and applicable specifications. This example is illustrative, not a universal ratio.
If the Check Points show large discrepancies between their surveyed coordinates and the coordinates extracted from the final model, the Surveyor should investigate the cause before accepting the deliverables.
Possible causes include poor image measurements, inaccurate ground coordinates, an unsuitable coordinate reference system, weak image geometry or inadequate control distribution.
Remember: a model’s fit to the GCPs does not, by itself, prove that the entire model is accurate. Independent Check Points provide a stronger test of positional accuracy.
5. What Makes a Good Ground Control Point?
Not every point on the ground is suitable for use as a GCP. The point must be measurable, identifiable and appropriately positioned.
A. Clear visibility
Targets should have strong contrast against their surroundings and be easy to identify in multiple photographs.
Avoid targets obscured by vegetation, shadows, vehicles, buildings or other objects.
B. Stable positioning
Targets should remain stationary throughout the survey. A target that shifts between the ground survey and drone acquisition can introduce errors into the processing.
C. Accurate coordinate determination
The coordinates must be measured using a method appropriate to the project’s accuracy requirements. The Surveyor should account for instrument performance, satellite conditions, datum, projection, geoid model where relevant and other sources of uncertainty.
D. Suitable distribution
Control points should be spread across the project area rather than clustered in one location.
Where practicable, distribute control around the perimeter and throughout the interior. Include appropriate elevation variation where terrain and project geometry require it.
A large project may need more carefully distributed control than a small, flat site.
E. Visibility in multiple images
A target should be visible and sharply defined in enough overlapping images to support reliable image measurement. Its suitability also depends on flight geometry and the processing software.
6. How Many Ground Control Points Do You Need?
There is no single number of GCPs that is appropriate for every drone survey.
The required number depends on several factors, including:
* The size and shape of the survey area.
* Terrain relief and elevation variation.
* The drone’s camera and positioning system.
* Image overlap and flight geometry.
* The required horizontal and vertical accuracy.
* Whether RTK or PPK positioning is used.
* Project specifications and the availability of independent Check Points.
A small, relatively flat site may require fewer control points than a large, mountainous or geometrically complex site. A long, narrow corridor may also require a different control arrangement from a compact construction site.
Rather than choosing a fixed number without considering the project, Surveyors should design a control network that meets the intended accuracy requirements and verify the result independently.
7. Can RTK or PPK Drones Work Without GCPs?
Yes, under suitable conditions, drones equipped with RTK or PPK positioning can produce georeferenced photogrammetric products with few or no GCPs.
RTK uses real-time positioning corrections, while PPK applies positioning corrections during post-processing. These methods can improve the accuracy of the camera positions recorded for each image.
However, RTK and PPK do not automatically guarantee that the final photogrammetric product meets the required accuracy.
Results depend on factors such as the quality of the GNSS observations, camera calibration, camera-to-antenna offsets, image geometry, coordinate reference systems and processing settings.
Even when GCPs are not required for model adjustment, independent Check Points remain valuable for verifying the output.
For projects with stringent accuracy requirements, difficult terrain or contractual specifications, a suitable ground control strategy may still be necessary.
The correct approach is to select the positioning and control method according to the project requirements, then verify its performance.
8. Common Mistakes to Avoid When Using GCPs
Surveyors should be alert to the following errors:
1. Poor distribution: placing all GCPs close together or along one edge of the project.
2. Incorrect coordinates: entering coordinates in the wrong datum, projection or units.
3. Poor target visibility: using markers that are too small, obscured or difficult to distinguish.
4. Insufficient image coverage: failing to capture targets clearly in enough photographs.
5. Using every surveyed point as control: leaving no independent Check Points to evaluate the model.
6. Ignoring vertical reference: mixing ellipsoidal heights and orthometric heights without appropriate conversion.
7. Relying only on processing residuals: assuming small GCP residuals prove the accuracy of the entire model.
8. Failing to document the survey: omitting coordinate systems, equipment, measurement methods and quality checks from the project records.
Avoiding these mistakes improves the reliability and defensibility of the final survey products.
9. Applications of GCPs in Professional Surveying
Ground control supports a wide range of drone mapping applications.
Topographic surveying: Helps reference terrain models, contour maps and other surface information to a defined coordinate system.
Construction monitoring: Supports comparison of site conditions with design information and monitoring of changes over time.
Land development: Helps integrate drone-derived mapping with engineering layouts, site plans and other spatial datasets.
Mining and quarrying: Supports georeferencing of terrain models and provides a basis for validating stockpile volume calculations and monitoring excavation progress.
Infrastructure mapping: Supports mapping of roads, drainage systems, utility corridors and other infrastructure features.
Environmental monitoring: Helps reference imagery used to assess erosion, vegetation changes, flood impacts and land-cover changes.
In every application, the appropriate accuracy depends on the project requirements, the quality of the control network and the processing methodology.
10. What Should Surveyors Report About GCP Accuracy?
A professional photogrammetric report should document more than the number of GCPs used.
Depending on the project specification, the report should include:
* The coordinate reference system, including the horizontal datum, projection and vertical reference.
* The method and equipment used to survey the control points.
* The distribution and identification of GCPs and Check Points.
* The number of images in which each control point was measured, where available.
* Processing residuals for the GCPs.
* Independent Check Point errors in horizontal and vertical directions.
* Summary accuracy statistics, such as RMSE, where appropriate.
* Known limitations, assumptions and quality-control procedures.
Root Mean Square Error (RMSE) is commonly used to summarise coordinate differences. However, it should be interpreted alongside the sample size, point distribution, error pattern and project acceptance criteria.
A report should not claim a particular level of accuracy simply because the software displays a small residual value.
Conclusion
Ground Control Points are a fundamental part of many accurate aerial photogrammetry workflows. They connect the photogrammetric model to surveyed ground coordinates and can help improve its positioning, orientation and scale.
However, successful drone mapping requires more than placing targets on the ground. It involves careful flight planning, reliable ground surveying, appropriate coordinate reference systems, good image geometry, sound processing and independent accuracy assessment.
For the modern Surveyor, understanding GCPs means understanding the relationship between aerial imagery and the real world.
The drone captures the data. The Surveyor establishes the control, evaluates the accuracy and ensures that the final map can be trusted.
Discussion Question
When conducting a drone survey, would you prefer to use traditional GCPs, an RTK/PPK workflow, or a combination of both? What factors would influence your decision?
Share your experience and join the discussion with the Young Surveyors Network FCT – Education Committee.
Learn | Connect | Build the Future
🌐 www.youngsurveyorsnetworkfct.ng
