RTK accuracy explained: why centimetre precision matters for your output
Centimetre accuracy is not a marketing tier. It is the difference between an output that can be used for cadastral or regulatory purposes and one that serves as a visual reference only.
An RTK receiver does not produce accuracy by itself. It measures the phase of the carrier signal from multiple satellites, and the position is resolved by comparing the arrival time of those signals against a reference station whose own position is known to centimetre level. That reference is the critical dependency, and it explains why RTK accuracy is described as relative rather than absolute. Two error sources dominate in practice: atmospheric delay, which grows with the square of the distance from the reference station, and multipath, where the signal reflects off nearby structures and the receiver resolves the reflected path instead of the direct one.
The practical consequence is that a single reference station has a working radius, typically 30 to 50 kilometres in open terrain and considerably less in urban or forested areas where multipath dominates. Beyond that radius, the corrections degrade and eventually the receiver falls back to standalone mode, which is where the familiar one to three metre accuracy of a consumer GNSS comes from. This is why professional survey work uses a base station at one end of the site and an RTK-capable rover on the aircraft, or a network of continuously operating reference stations.
For the buyer, the questions that separate a genuinely capable survey platform from one with an RTQ label are specific. Ask what the reference configuration is and whether the base station is included or needs to be supplied. Ask whether the platform supports a network RTK correction stream, which removes the need for an on-site base entirely and is the usual configuration for work across a wide area. Ask what the reported accuracy is measured against, and whether it is a horizontal root-mean-square error over a surveyed control network or a manufacturer figure from an open field with ideal satellite geometry.
One more practical point. RTK convergence takes time. A receiver needs a period of continuous tracking to resolve the integer ambiguities in the carrier phase, commonly under a minute in open sky but noticeably longer in poor visibility. On a mapping flight, that means the first point of each sortie is often lower quality than the rest. Survey workflows account for this by discarding the first points or flying a stabilising pass, and a buyer should ask whether the flight planning software does the same automatically.
