GNSS — the family of satellite navigation systems including GPS, GLONASS, Galileo and BeiDou — lets a receiver compute its position from satellite ranging signals. Uncorrected, that position is accurate to a few metres; surveying needs centimetres. Differential techniques such as RTK and PPK achieve exactly that, by measuring against reference stations whose positions are precisely known. They are the positioning backbone of modern field survey and drone mapping.
Why raw GNSS is metre-grade
Satellite signals cross the ionosphere and troposphere, ride on imperfect clocks and orbits, and bounce off nearby surfaces (multipath). Each effect injects error; together they leave a standalone receiver metres from truth — fine for navigation, unacceptable for boundaries and engineering.
The key insight of differential GNSS: most of these errors are shared across nearby receivers, so a reference station at a known point can measure the moment’s total error and hand nearby rovers the correction.
RTK and PPK in practice
Real-Time Kinematic (RTK) streams corrections from a base station or a network service (CORS/VRS) to the rover over radio or internet, resolving carrier-phase ambiguities to deliver centimetre coordinates live — the mode of stakeout and machine control.
Post-Processed Kinematic (PPK) logs raw observations on both ends and resolves them after the mission — no live link needed, more robust in difficult telemetry environments, and the favourite for drone camera positioning.
The details that bite
Fixed versus float matters: only an ambiguity-fixed solution is centimetre-grade; float solutions quietly drift to decimetres. Canopy, canyons and buildings degrade signal geometry. And heights deserve respect: GNSS measures ellipsoidal height, while projects need orthometric height via a geoid model — a distinction that has ruined many elevation datasets.
Professional practice logs solution quality, checks into known control, and records datum, epoch and geoid explicitly in every deliverable.