A drone survey proposal may promise a map at 2 cm per pixel. That figure describes image sampling; it does not establish the accuracy of every measurement made from the survey. For project teams buying or managing aerial mapping, understanding this distinction helps turn a resolution target into a usable specification.
Ground sampling distance, or GSD, belongs alongside the required deliverables, positional accuracy and acceptance checks. Choosing it in isolation can lead to unnecessary data collection or a map that looks sharp but lacks the evidence needed for engineering decisions.
Start with what the map must show
Identify the features your team needs to interpret: the outline of an excavation, visible stockpile boundaries, access roads or smaller exposed site features. Then identify the decision those observations will support. A general progress record and a detailed survey for engineering design are different assignments, even when both produce an orthomosaic.
Separate visibility from measurement. Recognising a feature in a photograph does not establish its position within the project tolerance. Likewise, a plan view alone cannot supply all the surface information needed for a volume calculation. Specify the map, point cloud or surface model that the task requires, and agree how its suitability will be demonstrated.
What GSD tells you
GSD expresses the ground distance between neighbouring pixel centres, usually in centimetres per pixel. A nominal value of 2 cm/pixel means that spacing is 2 cm under the assumed imaging geometry. A smaller number represents finer sampling. The Pix4D technical definition explains the relationship between GSD and image detail.
Pixel spacing is only part of image usability. Focus, motion blur, exposure and contrast affect whether a feature can be distinguished. A feature as narrow as one pixel is not automatically identifiable. The relevant question is whether it can be interpreted consistently in the actual imagery, at the locations that matter to the project.
Ask which resolution a supplier is reporting. Planned image GSD, an estimate from the captured images and the pixel size of an exported map are related but distinct. Selecting smaller output pixels does not create detail that was absent from the photographs. Our orthomosaic guide describes the wider checks needed for that deliverable.
Calculate an initial estimate
For a near-vertical camera viewing approximately level ground, a simplified estimate is GSD = H × S / (f × N). H is camera height above the ground, S is physical sensor width, f is actual focal length and N is image width in pixels. This follows the geometry in Pix4D's flight-height calculation.
If H is in metres and S and f use the same unit, the answer is in metres per pixel. Take a hypothetical camera with a 13.2 mm sensor width, an 8.8 mm focal length and a 6,000-pixel image width at 80 m above the ground. The calculation is 80 × 13.2 / (8.8 × 6,000) = 0.02 m/pixel, or 2 cm/pixel. This is an illustrative calculation, not a recommended aircraft, camera or operating height.
Use the actual focal length rather than its 35 mm equivalent. Check that sensor dimensions and image dimensions match the capture mode. A megapixel count alone cannot compare the expected GSD of cameras with different sensors and lenses. Camera tilt and terrain relief also limit how far a single nominal estimate describes the whole site.
Keep accuracy and point density in the specification
Image GSD, reconstructed point density and positional accuracy describe different aspects of a survey. Point density concerns the number of reconstructed points within a stated area or volume. Positional accuracy concerns agreement with reliable reference coordinates. Neither a dense point cloud nor a fine pixel grid proves that agreement.
Capture geometry, calibration, positioning and survey control all affect the result. The USGS guidance on UAS imagery calibration discusses geometric quality and validation. If the work requires demonstrated accuracy, ask for independent checkpoints that were not used to adjust the model, together with the assessment method and results.
Specify horizontal and vertical requirements for the intended use. A nominal 2 cm GSD should never be presented as a universal 2 cm maximum error. Site conditions and the measurement workflow must support the claim being made.
Account for coverage and terrain
With the same camera, flying closer to the surface reduces nominal GSD and the footprint of each photograph. Covering the same area with comparable overlap can therefore require more images and more work. Review whether the additional detail changes the decision your team can make before choosing the finest setting available.
Height above a take-off point is not necessarily height above the ground throughout a flight. Rising terrain brings the surface closer to the camera; falling terrain takes it farther away. These changes affect both sampling and effective overlap. A single project average can conceal less suitable coverage in a critical part of the site.
Pix4D's terrain guidance describes approaches such as terrain-following missions and separate flights. Their suitability depends on the aircraft, planning software, terrain information and operating conditions. Confirm the supported workflow before relying on it in a survey specification.
Turn the target into acceptance checks
Agree the checks before capture so that the field team and the people accepting the deliverables use the same criteria. A practical handover should answer the following questions.
- Are the required features visible across all priority areas?
- Are blur, gaps and difficult surfaces identified and documented?
- Are acquisition date, coordinate reference system and output resolution stated?
- Where accuracy is required, is the independent assessment included?
- For repeat surveys, are resolution and reference systems suitable for comparison?
Review a representative sample early when establishing a new workflow. Inspect the delivered data at useful working scales, and follow up on limitations before relying on measurements. A larger export or a sharper-looking display cannot replace missing observations or independent quality evidence.
The Maply platform brings imagery processing, map and model viewing, and measurement tools into one environment. To evaluate it against your requirements, request a demonstration with a representative area, the features you need to see and the decisions the survey must support. That gives the discussion a concrete basis beyond a single GSD number.
