Drones turned surveying from a days-long walk into a 30-minute flight — but only if the accuracy is real and, crucially, provable. RTK and PPK are what make that possible, and this guide covers how to turn a precisely-positioned drone into a survey a client will sign off on. It builds on RTK vs GNSS vs PPK and base stations & NTRIP.

A mapping drone deployed on a gravel bar in a mountain valley for a survey mission, with field gear laid out beside it
A survey drone deployed for glacier-terrain mapping in Alaska — RTK/PPK fieldwork happens in places ground survey can't easily reach. Photo: USDA Forest Service, Alaska Region (public domain, via Wikimedia Commons).

What "survey-grade" actually requires

Accurate camera positions are necessary but not sufficient. A defensible drone survey needs four things working together:

  1. Precise georeferencing — RTK or PPK camera positions (see the RTK vs PPK choice below).
  2. Good imagery — sharp photos at an appropriate ground sample distance, ideally from a mechanical-shutter camera.
  3. Sound geometry — enough overlap and a sensible flight pattern (the photogrammetry fundamentals).
  4. Verification — independent checkpoints that prove the delivered accuracy.

Skip any one and the others can't rescue you.

RTK/PPK vs ground control points

Traditionally you accurately surveyed ground control points (GCPs) — marked targets on the ground — and the software used them to georeference the map. It works beautifully but the GCP walk is often the dominant field cost. RTK/PPK shrinks or removes it by giving each photo an accurate position directly. The nuance every professional keeps:

  • RTK/PPK replaces most GCPs for georeferencing — huge time savings.
  • It does not replace checkpoints. Keep a few independently-surveyed points you don't feed into processing, purely to measure the final error. That's the difference between "we used RTK" and "here is our verified 3 cm accuracy."

RTK or PPK for surveying?

For mapping, only the photo positions must be accurate, not the flight itself — which is exactly PPK's sweet spot. PPK logs raw data and computes positions after landing, so a correction-link dropout mid-flight can't wreck the job; many survey operators default to it for that robustness. Use RTK when you also need the drone to fly precisely (tight repeat lines, obstacle-close work) and your correction link is solid. Plenty of aircraft support both — record raw data even when flying RTK, so you have a PPK safety net.

Planning the flight

  • GSD first: pick the ground sample distance your accuracy needs (finer GSD → lower flight, more images, longer flights). Plan horizontal accuracy of ~1–2× GSD, vertical ~2–3×.
  • Overlap: 75/75% front/side is the modern default; raise it over uniform or vegetated terrain.
  • Constant height above terrain, not take-off — follow the ground on sloped sites.
  • Camera events: make sure the precise position is stamped at the exact moment of exposure (hot-shoe / PPS sync) — the time-sync discipline that mapping payloads live or die on.
The workflow mistakes that ruin surveys

Trusting RTK without any checkpoint (you can't prove the result); a self-averaged base when the job must tie to a national grid (tight relative accuracy, wrong absolute position); mismatched coordinate systems and geoid models (centimetre data, metre-level blunder); and testing the workflow on the client's job instead of before it. None are about the drone — they're about discipline.

Where LiDAR comes in

Photogrammetry with RTK/PPK is superb on hard, textured surfaces. Over vegetation, or for wires and bare uniform ground, you'll want LiDAR instead — and there RTK/PPK matters even more, because every laser point's accuracy depends directly on knowing the sensor's position and attitude at the instant of the pulse.

Frequently asked questions

How accurate is RTK drone surveying?

With a good RTK or PPK workflow and proper checkpoints, drone surveys reach a few centimetres horizontally and roughly 1.5–2× that vertically — genuinely survey-grade for many applications. The final accuracy depends on ground sample distance, camera quality, georeferencing and verification, not just on having RTK.

Do I still need ground control points with an RTK drone?

You need far fewer, and sometimes none for relative accuracy, but you should always keep at least a few independent checkpoints to verify the result. RTK/PPK gives accurate camera positions; checkpoints prove the final map is actually as accurate as you claim, which is what a client or auditor wants.

What accuracy do I need for a drone survey?

It depends on the deliverable. Volume calculations tolerate a few centimetres; construction stakeout and legal boundaries need tighter, verified accuracy tied to a coordinate system. Define the required accuracy first, then design the flight, georeferencing and checkpoints to prove you met it.