RTK, PPK & Ground Control, Explained

When someone says a map is accurate to two centimeters, ask: compared to what? Relative accuracy is how tight the map is within itself, one point measured against another. That drives most of what you care about: stockpile volumes, cut and fill, distances, grade changes. Absolute accuracy is how well the whole map lines up with true real-world coordinates, which matters when you overlay a design file, compare flights months apart, or tie into records.

Here's the nuance that trips people up: RTK and PPK deliver centimeter accuracy relative to their base station almost automatically. But if that base isn't tied to a known point or a network correction, the entire job can sit a meter or more off in true position while every internal measurement looks perfect. A popular base station powered on cold is only accurate to about 1.5 m horizontal and 3 m vertical on its own. Setting the reference up correctly is the part you're actually paying a professional for.

Ground control points are the original method: lay four or more marked targets across the site, measure each one's coordinate on the ground, and the software stretches the map to match. It reaches roughly 1 to 3 cm, but the accuracy is bought with field labor on every single flight, and laying a full grid can take 30 minutes to over 2 hours per site.

RTK (Real-Time Kinematic) streams corrections to the drone live, from a base station or a cell-network service, and stamps every photo with a centimeter position as it's captured. Typical results run 1 to 3 cm horizontal and 2 to 5 cm vertical, and the pilot can confirm quality before leaving your site. Its weak point is the live link: lose the radio or cell connection mid-flight and the fix degrades with no way to recover it afterward.

PPK (Post-Processed Kinematic) records raw satellite data during the flight and runs the correction math afterward in software, typically landing at 1 to 2 cm horizontal and 2 to 3 cm vertical. Because nothing depends on a live connection, a signal dropout doesn't ruin the flight, which makes PPK the most forgiving choice on remote, large, or obstructed sites.

There's a difference between software saying a map is accurate and a map checked against the ground. The professional standard is a small set of independent checkpoints: extra points measured in the field but held out of processing, then compared against the finished map. That comparison, reported as measured error, is what separates a defensible number from a good-looking guess.

This is also the honest version of 'no ground control needed.' RTK and PPK remove the dense grid of targets that creates accuracy, but a few checkpoints that verify accuracy stay on every professional job. A field test by Pix4D makes the stakes clear: the same flight hit about 2.5 cm with precise positioning and drifted roughly 3.6 m vertically in plain GPS mode, about a hundredfold difference you'd never see just looking at the map.

RTK is the default for connected, fast-turnaround work: construction sites with good cell coverage, roof and facade capture, and any job where you want quality confirmed on-site before packing up. PPK wins on large, remote, or obstructed sites, and on volume-critical work where consistent vertical accuracy protects the cut/fill and stockpile numbers. Ground control alone still makes sense on small, open, one-off sites where the walk is short.

In practice the methods stack: a common professional setup flies RTK for live positioning while also logging raw data, so the flight can be re-processed PPK-style if the link dropped. You don't need to memorize any of this; a provider who explains which method your site calls for, and why, is showing you their work.