PilotLedger emblem PilotLedger
← Back to Blog
PL

What Is RTK? Drone Mapping Explained

If you have never set a base station, asking what is RTK is fair. The acronym shows up on aircraft listings, quote templates, and client emails as if everyone already knows it. Most people who fly their first mapping job do not.

RTK stands for Real-Time Kinematic. In RTK drone mapping, the aircraft is the rover. It uses live GNSS correction data so the positions written to each photo are improved while you fly, not after you get home.

I fly mapping work commercially out of Springdale, Arkansas. This is the field explanation I wish I had the first time someone said to turn RTK on. Pilot Ledger is not an RTK rover. The question is what the correction is doing.

What the correction is for

A GNSS receiver without a correction link, including the one in a typical drone, is usually good to a few meters. That is enough to stay on a flight line. It is not enough if the client will drop your orthomosaic onto a civil set and expect a curb to land on a curb.

The satellites are the same. The error sources are the same: clocks, orbits, atmosphere. A second receiver sitting still — a base — sees those errors too. RTK sends that receiver's corrections to the moving rover fast enough that the rover can apply them on the spot.

Those positions can be centimeter-level only when the correction link is good, the rover reports a fixed solution, and you still take independent check shots. RTK is a live correction method. It is not a stamp that the map is finished.

Two ways the rover gets corrections

Most mapping days use one of two paths.

A local GNSS base

You set a GNSS receiver on the ground and leave it there for the flight. That is the base. The drone is the rover. The base sends correction messages over a radio, or sometimes a local network, while you fly.

The point the base sits on matters. If you occupy a known point — a monument, control you already hold, or a coordinate the surveyor gave you — the rover's positions are tied to that point in real time. If you occupy an unknown point, the flight can still hold together, but the whole map is only as good as the coordinate you later assign to that base. You can occupy now and adjust later. You cannot skip the known point and call the map georeferenced.

If you have never set a base, the sequence is simple: find open sky, set the tripod, measure the antenna height, start the base logging and transmitting, confirm the rover is hearing it, then fly. The base is a still receiver that tells the moving receiver how to correct.

NTRIP, or network RTK

NTRIP is Networked Transport of RTCM via Internet Protocol. In plain English, it is a correction stream that arrives over cellular from a caster — a server that publishes corrections from a network of reference stations. That setup is often called network RTK.

You do not set a tripod. You log into a caster with a mountpoint and credentials, and the rover consumes that stream the same way it would consume a local base. The trade is convenience for a live cell link. If the site has no data, NTRIP does not run. If the mountpoint is far from the job, or the stream drops, you are not in a finished RTK solution even if the aircraft is still flying.

A local base and NTRIP are two delivery methods for the same idea: live corrections to the rover. They are not two different kinds of accuracy.

Fixed versus Float

Your controller or aircraft status will usually show Fixed or Float. That one word is the status you need on site.

Fixed means the rover has resolved the integer ambiguities in the carrier-phase measurements. The unknown number of whole radio cycles between each satellite and the antenna has been decided. That is the solution you want before you treat a photo position as survey-useful.

Float means corrections are arriving, but those integers are still ambiguous. The rover has a better guess than raw GNSS. It does not have a finished survey position. Do not keep flying a map you intend to overlay on a civil set while the status sits on Float and hope processing will clean it up.

You do not need the linear algebra. Watch the status. If it is not Fixed, you do not have RTK in the sense the client thinks you do.

When the live link drops

A radio fade behind a berm, a cell dropout in a valley, a caster that stops: any of those lose the real-time link. The aircraft can keep flying. The positions stop being corrected in real time. That is why PPK exists: corrections can also be applied after the flight.

If you are flying RTK, treat a lost link as a lost correction. Restore the link, confirm Fixed, then continue. Logging that the aircraft is RTK-capable is not the same as a flight that stayed Fixed.

What RTK does not replace

RTK does not magically replace check shots or a known point.

The correction is relative to the source you used. A local base on a bad coordinate produces a tight map in the wrong place. An NTRIP stream in a datum the civil set does not use produces a clean overlay that misses. Absolute accuracy is where the map sits on the Earth; relative accuracy is how well the map holds together with itself. RTK can help both. It does not certify either.

If the deliverable will be overlaid on a civil set, take independent check shots — GCPs or checkpoints you did not use to steer the solution — and report how the map sits on those points. A fixed RTK flight with no checks is a claim. A fixed RTK flight with checks is a map you can stand behind.

The known point still matters. The coordinate system still matters. The vertical datum still matters. Those are survey decisions, not aircraft settings.

What to confirm before you take off

None of that requires you to be a geodesist. It requires you to treat RTK as a live correction link, not a mode you toggle once and forget.

If a site or a status in the field still is not clear, ask the follow-up on Ask Mav at pilotledger.com.