C‑DRONE
Top-down orthophoto-style view of farmland parcels captured by drone

C-DRONE GUIDE · 26 AUGUST 2026

Ground sampling distance (GSD) in drone photogrammetry: choosing it project by project, price

Two drone photogrammetry missions can produce radically different deliverables for the same site, with neither one being "wrong": it all comes down to the ground resolution requested, or GSD (ground sampling distance) — the real, on-the-ground size that each image pixel represents. A municipal orthophoto at 8 cm GSD is plenty to place a building on a local land-use plan; it will never reveal a 2 mm crack on a facade. Conversely, flying low enough to reach 3 mm GSD over a hectare of roof multiplies the number of images, the processing time and, often, the price by five. This guide explains how GSD is calculated, what resolution to request depending on the use case, and how it feeds through into a 2026 quote (excl. VAT).

Published on 26 August 2026, reviewed on 27 August 2026 — regulations in force as of August 2026.

What GSD is, and what it depends on

Ground resolution, or GSD (ground sampling distance), is the real on-the-ground distance covered by a single pixel of the captured image. A GSD of 2 cm means each pixel corresponds to a 2 cm square on the ground: two objects less than 2 cm apart will appear as a single pixel, indistinguishable from one another in the image. The smaller the GSD, the finer the usable level of detail — and the more the file size, flight time and processing time increase in proportion.

Three parameters set the GSD of a shoot: flight altitude (GSD doubles when altitude doubles, for the same sensor and focal length), the lens's focal length (a longer lens narrows the field of view and sharpens the GSD without flying lower) and the sensor's pixel size, which is specific to each onboard camera model. For the same mission budget, a pilot equipped with a newer sensor can therefore achieve a finer GSD at the same altitude than with older equipment — which is why two quotes for the same flight sometimes list different resolutions.

A useful benchmark: at 50 m altitude, most professional large-sensor drones produce a GSD of around 1 to 1.5 cm; at 120 m, the open category's ceiling, it rises to 2.5-4 cm depending on the sensor. Flying below 30 m sharpens the resolution further, but reduces the area covered per image just as much and requires flying much closer to obstacles.

Which GSD for which use case

The resolution needed depends entirely on what the deliverable must show, not on a universal standard. A few benchmarks seen in professional use:

A landmark study by James and Robson, published in 2012 in the Journal of Geophysical Research: Earth Surface, showed that structure-from-motion/multi-view-stereo 3D reconstruction — the technique used by every drone photogrammetry software today — typically reaches a precision ratio better than 1:1000 relative to the shooting distance, meaning centimetre-level accuracy over measurements of several tens of metres (see the study on Google Scholar). Applied over a year to a 50 m coastal cliff, seven successive surveys measured an average retreat of 0.70 ± 0.05 m per year — a variation that a coarser GSD would simply have buried in measurement noise.

GSD, overlap and ground control points: the trio behind accuracy

A fine GSD guarantees nothing on its own. Two other settings determine the quality of the final model. Overlap between consecutive images — typically 70 to 80% along the flight line and 60 to 70% between flight lines for a standard survey, more over rugged terrain or a complex facade — lets the software reconstruct each point on the ground from several angles; insufficient overlap creates holes or distortions in the model, however fine the pixels are. Ground control points (GCPs), surveyed with centimetre-grade GPS and spread across the site, then anchor the model in a real coordinate system rather than a relative one.

A study by Agüera-Vega, Carvajal-Ramírez and Martínez-Carricondo, published in 2017 in the journal Measurement, precisely quantified this role: by varying the number of ground control points used on the same UAV survey site, the authors showed that the accuracy of the resulting digital terrain model depends as much on the GCP network as on the resolution of the images themselves, with the best configurations reaching a map scale of 1:150 and a 15 cm contour interval (see the study on Google Scholar). A 5 mm GSD without enough well-distributed GCPs can therefore produce a model that is very detailed locally but shifted by several centimetres overall — not good enough for legal use or a utility as-built survey. Conversely, an onboard RTK or PPK GPS (see our guide to drone RTK or PPK) substantially reduces, without eliminating, the need to multiply ground control points.

GSD and the flight framework: what limits the choice

The finest possible GSD is not always the most relevant one, nor the most achievable. Flying low to sharpen resolution mechanically brings the aircraft closer to obstacles, to people and sometimes to an unconsented overflight of third-party property: above 120 m, the open category's ceiling, an obstacle exemption and a specific case file become necessary (see our guide to flying a drone above 120 m), while below a few dozen metres, proximity to the ground, fences or power lines calls for more cautious flying and sometimes a ground observer. A longer-focal-length sensor or an optical zoom often sharpens the GSD without flying lower, particularly for inspecting tall structures (pylons, chimneys, building facades) where getting physically closer is ruled out anyway.

The GSD should therefore be chosen upfront, with the pilot, based on the deliverable's actual use — never by default, or "as fine as possible". An oversized resolution relative to the need adds weight to the flight, the processing and the bill with no benefit to the client; an undersized one means having to fly again.

How GSD feeds through into the price

For the same area, the GSD requested is one of the main drivers of a photogrammetry mission's price, even before the choice of final deliverable. A finer GSD requires a lower altitude, hence more flight lines and images to cover the same area, longer computer processing time, and often extra ground control points so that geometric accuracy keeps pace with visual resolution.

Orders of magnitude observed in 2026, excluding VAT, for a one-hectare photogrammetric survey:

To get an accurate quote, state the deliverable's actual use in your request (land-use plan, building permit, crack monitoring, digital twin) rather than a GSD in centimetres: the pilot translates that need into the right altitude, focal length and overlap. Request a quote, specifying the area, the intended use and the expected deliverable format (orthophoto, point cloud, 3D model).

Frequently asked questions

What GSD should I request for a simple orthophoto or photo shoot?

Between 5 and 8 cm is more than enough to place a building on a local land-use plan or produce a clear overview of a site. Going finer adds nothing for this use and only lengthens the flight and processing time for no benefit.

Why does a very fine GSD cost noticeably more?

Because it requires flying lower or using a longer focal length for the same area, meaning more flight lines and images to process, and it almost always comes with an extra network of ground control points so the model's geometric accuracy keeps pace with the pixels' visual sharpness.

Does a fine GSD alone guarantee the accuracy of a survey, for example for a utility as-built survey?

No. GSD only measures the visual detail of pixels; absolute geometric accuracy depends mainly on the overlap between images and on ground control points surveyed with centimetre-grade GPS. A Class A buried-utility survey requires a dedicated GCP network on top of a fine GSD — see our guide to buried-utility as-built surveys.

Request a free quote

Put it into practice

Also worth reading