Drone Surveying & Photogrammetry: The Complete Guide

Two technologies dominate drone surveying, and they solve different problems. Photogrammetry reconstructs 3D geometry from overlapping photographs — cheap, colour-rich, and excellent for surfaces and orthomosaics. LiDAR fires laser pulses and measures their return — more expensive, but it sees through vegetation to the bare earth.

For most jobs — earthworks, site documentation, stockpile volumes — photogrammetry is the pragmatic choice. Reach for LiDAR when vegetation hides the ground and you need a true bare-earth surface. The resulting point clouds read into any standard CAD or GIS package as LAS, LAZ, PLY or XYZ files.

Ground Sample Distance (GSD) is how much ground each pixel covers — the master quality dial of any photogrammetric survey. A GSD of 2 cm/pixel means one pixel equals 2 cm on the ground; finer GSD means more detail and, as a rule of thumb, better achievable accuracy.

GSD is governed by a simple relationship:

GSD = (sensor pixel size × flight height) / focal length

The practical lever is flight height: fly lower for finer GSD, higher for more coverage per battery. A useful planning anchor is that final accuracy is often roughly 1–3× the GSD when ground control is good. So if a client needs 3 cm accuracy, you plan a flight whose GSD is around 1–1.5 cm and you do not cut corners on control.

Photogrammetry needs every point on the ground to appear in many photos. The standard is 70–80% front overlap (along the flight line) and 60–70% side overlap (between lines). Increase both to 80–85% over vegetation, water, or uniform surfaces where the software struggles to match features.

A solid flight plan also fixes the camera angle and pattern: nadir (straight down) grids for terrain and orthomosaics; oblique passes for vertical structures and 3D models. Plan the mission so the whole area is covered with consistent overlap, and check the GNSS window for the day — even drones benefit from good satellite geometry for their onboard receiver. Our GNSS Mission Planner shows the satellite count and PDOP for your site and time.

A point cloud with no georeferencing is just a pretty shape. To place it correctly and accurately in the real world, you have two approaches — and the best surveys often combine them.

Ground Control Points (GCPs)

GCPs are marked targets on the ground whose coordinates you measure precisely with GNSS or a total station. The photogrammetry software ties the model to these points. Five to ten well-distributed GCPs — including some near the edges and at varied elevations — transform a relative model into a survey-grade deliverable. GCPs remain the gold standard for verifiable accuracy.

RTK / PPK drones

A drone with an onboard RTK or PPK receiver records the precise position of every photo as it is taken, drastically cutting the number of GCPs needed — sometimes to zero for relative accuracy. PPK is especially robust because it post-processes against a base or CORS station and never depends on a live radio link mid-flight.

Best practice: use RTK/PPK for efficiency and place a handful of independent checkpoints — GCPs you do not feed into processing — to measure and certify the accuracy you actually achieved.

Vertical accuracy is always the weaker axis in photogrammetry, typically 1.5–2× worse than horizontal — which is exactly why checkpoints and good GCP vertical distribution matter so much. Remember too that your drone outputs ellipsoidal height from GNSS; converting to elevation above sea level needs a geoid model, as covered in the Coordinate Systems guide.

• Orthomosaic — a single, geometrically corrected aerial image you can measure directly. The base layer of most deliverables.
• Point cloud — millions of XYZ (and often RGB) points; the raw 3D record of the surveyed surface.
• Digital Surface Model (DSM) — elevation of everything, including buildings and trees.
• Digital Terrain Model (DTM) — bare earth with objects removed; the basis for contours and earthworks.
• 3D mesh — a textured surface model for visualisation and inspection.
• Contours & volumes — derived products for design and stockpile measurement.

State the coordinate system of every deliverable explicitly — look up the right EPSG code in the EPSG Explorer and confirm the country's official datum in the Country Profiles before export.

1. Define the accuracy target with the client, then back-calculate the GSD and flight height you need.
2. Check airspace and regulations for the country and site. Rules differ widely — always fly legally.
3. Plan the flight with correct overlap and pattern, and verify the GNSS window in the Mission Planner.
4. Lay and measure GCPs/checkpoints with RTK GNSS, tying into a known datum.
5. Fly in stable light and low wind; avoid harsh shadows and reflective water where possible.
6. Process and georeference, then validate against the independent checkpoints.
7. Export in the agreed EPSG code and document the achieved accuracy in the report.

Every acronym above is defined in the surveying glossary, and you can compare drone and sensor makers in the manufacturers directory.