A drone photograph can look impressively sharp on a screen and still be unsuitable for the measurement someone wants to make. The missing piece is often not megapixels alone, but the scale represented by each pixel on the ground. In mapping and photogrammetry, that idea is described by Ground Sampling Distance, usually shortened to GSD.
GSD is a useful planning term because it connects camera geometry with the detail visible in the captured image. It also prevents a common misunderstanding: a small GSD does not, by itself, guarantee that a map is accurately positioned or that every feature can be measured to the same precision.
What GSD actually measures
PIX4D defines GSD as the ground distance between the centres of two consecutive pixels. In practical terms, a GSD of 5 centimetres means that one pixel represents a ground distance of about 5 centimetres in the image. A GSD of 30 centimetres represents a much larger ground distance for each pixel, so smaller details are less visible.
The difference is not merely a label in a processing report. If a square pixel represents 5 cm by 5 cm on the ground, its footprint is about 25 square centimetres. At 30 cm by 30 cm, the footprint is about 900 square centimetres. Those examples describe the sampling scale; they do not say that an object of that size will always be detected, classified or measured correctly.
A pixel is a sample of image information, not a guarantee that the software can distinguish every object that occupies it. Contrast, focus, motion, surface texture, lighting and image processing still matter.
Why altitude and camera design change the number
GSD depends on the distance between the camera and the surface, together with the camera’s imaging geometry. PIX4D identifies flight height or distance to the terrain, image width, sensor width and focal length as the relevant inputs.
That is why changing altitude changes the ground area represented by each pixel, even when the camera itself has not changed. Flying closer generally produces a smaller GSD and more visible ground detail. Flying higher generally increases GSD, but can cover a larger area with fewer images and reduce acquisition and processing demands when the project does not need fine detail.
There is no single best GSD for every mission. A broad site overview, a stockpile survey and a detailed inspection have different information needs. The appropriate target should be chosen before capture, not selected after processing simply because a report displays a small number.
GSD is not the same as accuracy
This distinction is important. GSD describes the image’s sampling distance on the ground. It does not tell you how close the reconstructed coordinates are to their true positions.
Esri distinguishes relative accuracy from absolute accuracy in drone mapping. Relative accuracy describes how well objects relate to one another within the output. Absolute accuracy concerns how closely the output is placed in the real-world coordinate system. Esri notes that ground control points, or suitable RTK or PPK workflows, can be part of achieving stronger absolute positioning.
A map can therefore have a fine GSD and still be shifted, tilted or distorted in relation to surveyed control. Conversely, a project with a coarser GSD may be entirely adequate for an overview where exact small-feature measurement is not required.
Before accepting a deliverable, ask two separate questions: what ground detail can the imagery sample, and how has the position of the resulting map been checked?
The number may vary across one project
A planned flight height does not make every image’s GSD identical. PIX4D notes that terrain elevation differences and changes in camera angle can cause GSD to vary, even when the aircraft maintains a nominally constant height. Processing software may report an average value for an orthomosaic, but an average does not mean every pixel in every source image was captured at precisely that scale.
This matters over sloping terrain, tall structures and sites where the camera is not always pointing straight down. A flight plan that ignores terrain and object height can produce images with different scales and different amounts of detail.
Esri’s Drone2Map guidance also stresses maintaining a consistent distance to the terrain or object for effective 2D and 3D modelling. Where a project needs multiple scales, Esri recommends treating the scales as independent acquisition and processing projects rather than assuming they are interchangeable.
A smaller GSD can cost time and coverage
It is tempting to treat the smallest possible GSD as the automatic goal. That can be inefficient. Lower flight height may require more flight lines and more images to cover the same area. More images can mean additional battery changes, storage, transfer time and processing.
The correct trade-off depends on the smallest feature that matters, the size of the site, terrain, lighting, available flight time and the quality requirements of the final product. PIX4D explicitly notes that flying higher with a larger GSD can reduce acquisition time, battery use and processing time when the project does not require very detailed reconstruction.
Image overlap is another part of the decision. Esri explains that overlapping images allow photogrammetry software to match common features and build 3D information. Its guidance gives different overlap needs for different scene types; repetitive or featureless surfaces can require more overlap and may still be difficult to reconstruct reliably. A small GSD cannot compensate for missing overlap or a surface with too few identifiable features.
A practical planning checklist
Before a mapping flight, define the deliverable and work backwards:
- Identify the smallest feature the client or project actually needs to see.
- Choose a target GSD appropriate to that feature and the intended output.
- Check the camera’s sensor and focal-length specifications rather than relying on a marketing megapixel figure.
- Plan height and distance with terrain and tall objects in mind.
- Set suitable front and side overlap for the surface, pattern and reconstruction method.
- Decide how positional accuracy will be checked, including whether ground control or RTK/PPK is needed.
- Review the processing report for variation, gaps and weak areas instead of relying on one average number.
A useful GSD target is one that supports the job without creating unnecessary coverage and processing costs. It is a planning measure, not a certificate of survey accuracy.
Sources and editorial notes
- PIX4D, “Ground sampling distance (GSD) in photogrammetry”: https://support.pix4d.com/hc/en-us/articles/202559809
- Esri, “Generating high quality outputs from drone imagery”: https://www.esri.com/arcgis-blog/products/drone2map/3d-gis/generating-high-quality-outputs-from-drone-imagery
Sources checked: 2026-11-16. This article explains general photogrammetry concepts. It does not report a field survey, validate a particular aircraft or software configuration, or provide a guaranteed accuracy figure. Project requirements and applicable flight rules must be checked independently.
Leave a Reply