• Two types of accuracy — Relative accuracy is about reliable measurements within the dataset, while absolute accuracy is about placing the map or model correctly in real-world coordinates.
• Be careful with the word “survey” — A highly accurate drone deliverable can support survey-grade work, but that does not automatically make it a legal land survey.
• Know the three main accuracy tools — RTK, PPK, and Ground Control Points each help improve positioning accuracy, but they do so in different ways and at different stages of the workflow.
• Match the workflow to the project — RTK works in real time, PPK applies corrections after the flight, and GCPs help anchor and verify the dataset when higher-confidence absolute accuracy is needed.
• Treat accuracy as a set of decisions — The right approach depends on the site, connectivity, legal context, client needs, and how much confidence the final deliverable must support.
  

Now that you know the different types of mapping and modeling deliverables as well as the drone systems and additional equipment required to produce them, it’s time to talk about accuracy.

I call this lesson the art of accuracy, and that’s because you’re not just flying a drone mission — you’re deciding what level of precision is needed, what tools are appropriate, and how to produce a deliverable you can stand behind.

In this lesson, you’ll learn about relative vs. absolute accuracy. You’ll also learn the three main tools used to improve accuracy in drone mapping — RTK, PPK, and Ground Control Points. You’ll learn how each of these tools work and how to choose the right approach (or combine them) to produce reliable, measurement-grade results.

OK, let’s jump in.

Relative vs. Absolute Accuracy

First, let’s start by unpacking the word accuracy and how there are a couple of ways to think about what it means to produce a map or model that’s accurate.

First is relative accuracy. Relative accuracy is about how consistent and accurate the measurements are inside your map or model. If you measure the distance between two points, the area of a surface, or the volume of a stockpile, are those numbers reliable within that dataset? In many real-world jobs, this is what the client cares about most. They want dependable measurements for their particular site. They don’t necessarily care if the map’s exact placement on Earth is accurate.

Placement on Earth? What do I mean by that? Well, you have relative accuracy, and then you have absolute accuracy.

Absolute accuracy is about how well your map aligns to a global coordinate system. In plain speak, it’s the question of whether your deliverable is correctly placed on the planet, in the right latitude, longitude, and elevation.

While some clients only care about the relative measurements within your data (relative accuracy) other clients need your map or model to plug into a much bigger picture.

Let’s say a client wants to overlay your orthomosaic on top of existing GIS layers, combine it with CAD plans, compare it against last month’s survey data, or use it to stake out a specific location in the field. Then your dataset can’t just be floating around on Earth. It has to line up correctly with everything else.

So, with relative accuracy, you can have a map with the right shape and reliable measurements within it, but the whole thing might be shifted ten feet to the east of its actual location on planet Earth. For some jobs, that shift doesn’t matter. But for other jobs, that shift is a dealbreaker.

That’s where absolute accuracy matters.

Survey-Grade Data

Now, before we go any further, I want to call out a word that gets used casually but can carry real legal weight: “survey”.

You’ll hear people say “drone survey” all the time. It’s common language in the industry. But in many situations, the word “survey” refers to a specific licensed service with actual legal ramifications.

A drone map can be highly accurate, and it can even be at the level of “survey-grade” in the sense that it meets tight accuracy tolerances. But that does not automatically make it a legal land survey. When a project requires a stamped survey, that work needs to be done under the right professional authority.

So a good habit as a drone pilot is to be careful with how you describe your deliverables.

That said, drone mapping is still incredibly valuable for many projects that do not require formal survey deliverables because it can reduce costs, speed up decision-making, and provide measurement-grade information.

So I just wanted to explain that distinction because I have seen some drone pilots oversell or not really understand what survey-grade data is when they’re pitching their drone mapping and modeling services to clients.

Tools to Improve Accuracy

Okay, so if accuracy matters this much, how do we achieve it or improve it?

Most of the time, it comes down to how well you can control the positioning of the dataset. You see, when you take photos with your drone, your drone tags each photo with GPS coordinates. And those coordinates are good, but not perfect. If you want to tighten that up, there are three different ways to do it: RTK, PPK, and Ground Control Points.

Let’s start with the two that sound the most similar: RTK and PPK.

RTK stands for Real-Time Kinematic. PPK stands for Post-Processed Kinematic.

Both approaches have the same goal — to improve the accuracy of the GPS position data for each image. The difference between them is when that correction happens.

With RTK, Real-Time Kinematic, the correction happens during the flight, in real time. When RTK is working properly, your images are captured with near centimeter-level photo positioning baked in, so after capturing the data, you can move straight into processing, which is the stitching together of those images, without needing to do an extra correction step.

The way this works in basic terms is the drone has an RTK module on it — this module is built into some drones, and if not, then you can attach the module to the drone. This RTK module is a receiver that gets correction data in real-time from a known reference source. The reference source is usually a GNSS (Global Navigation Satellite System) base station that you’re setting up on site, or it’s a nearby network reference source that you’re tapping into. If you’re using a network reference source, you want the reference station to be reasonably close, ideally within a few miles of the site. The farther away it is, the more your accuracy can degrade.

The big requirement with RTK, whether you’re using the GNSS base station or the network service, is that it depends on a continuous connection during the mission. If that correction link drops, you might lose the benefit of RTK for part of the flight, and the results may be less consistent.

Let’s dive deeper into these two options — the GNSS base station and the network service.

First up is the GNSS base station. This is basically a receiver that you set up on site in a fixed position, and it broadcasts correction data that your drone uses during the flight.

As with anything, you have a range of products to choose from, depending on your needs.

On the more budget-friendly end, you’ll see DIY-style RTK kits built around GNSS boards from companies like SparkFun. These can absolutely work, but they require a bit more setup knowledge, and your results depend on things like antenna quality, placement, and how clean your base location is.

Then you have more “field-ready” receivers that are designed for mapping workflows and are easier to deploy consistently, like Emlid’s Reach series.

And at the higher end, you’ve got professional survey-grade systems from companies like Trimble, which are built for maximum reliability in tough conditions, but they come with a much higher price tag and are typically used by surveyors and engineering firms.

The point is not that one is always “best” but that the tool should match your accuracy requirement, your workflow, and the stakes of the project.

The second option for your reference source, and I mentioned this earlier, is to skip setting up a physical base station and instead subscribe to a network RTK correction service. In that workflow, your RTK-enabled drone system connects over the Internet to a corrections provider, which delivers the appropriate corrections to you in real time. Point One is an example of a commercial service in this category.

As I mentioned earlier, RTK requires a continuous connection, so what if that doesn’t work in your situation? That’s where PPK (Post-Processed Kinematic) comes in.

With PPK, you still rely on a known reference source, but you don’t need a live correction link during the flight. Instead, the drone’s onboard GNSS system records raw positioning and timing data as images are captured, while a separate GNSS reference receiver on the ground records raw satellite data over the same time period. After the flight, software compares those datasets and calculates corrected positions for the images. In other words, the corrections are applied after the fact rather than being delivered live during the mission.

One of the reasons people like PPK is its reliability. It’s a strong option in areas with poor connectivity, and it can be a great backup plan when RTK isn’t stable. It’s also a workflow that many operators trust for high-accuracy work because you can validate and control the correction process after the mission.

What I'm talking about is Ground Control Points, or GCPs. A Ground Control Point can look a few different ways, but in all cases it's a visible marker that is big enough to see from the air, and you or someone else places it on the ground in your flight area.

There is a methodical way to place them on the site — ideally, around the edges, across elevation changes, and near key areas of interest.

So you or someone else places the GCPs on the site, and then they mark the coordinates of each GCP with survey-grade positioning equipment, like a GNSS receiver. So you know the exact GPS coordinates of each GCP. Think of it like X marks the spot.

And here’s the thing. Often, it’s a licensed surveyor who is placing the GCPs and recording their location. You could do this yourself, but on larger projects, you’re working with someone else on the setup of GCPs.

So how many would you need? A small site might use 3 to 5 GCPs, a medium site maybe 5 to 10, and larger projects may require 10 to 15 or more. This can take a lot of time to do properly.

As your drone flies the mapping mission, it’s picking up your GCP targets on the ground, and those GCPs are hopefully showing up in multiple images.

When you get to processing, you tell the software, “These targets correspond to these known coordinates.” That gives the software fixed reference points it can use to properly georeference the map and improve absolute accuracy.

And one more quick thing I’ll say about GCPs is that they can be used in two different ways. You can use them as your primary accuracy method, or you can add a few as checkpoints alongside RTK or PPK to verify the accuracy of your results.

Now, we have three different techniques here, so what’s the “best” approach?

Some projects use only RTK. Some use PPK because they don’t have connectivity in the field. Some use only GCPs. Some projects use GCPs in addition to RTK or PPK, like I explained earlier.

As you think about applying each of these tools to your own mapping projects, here are some considerations.

And we've covered some of this already, but it's helpful to see it charted out like this.

First, when it comes to your drone, if you're just using GCPs, it doesn't matter what drone you're using, but if you're using RTK or PPK, you need an enterprise-level drone with those capabilities.

Second is your other equipment. With GCPs, you need the visible GCP targets that you're placing around the site. You also need a means of measuring those GCPs. Most people are using a GNSS rover, where you're placing that rover on top of each GCP. And that rover needs to get correction data, either from a GNSS base station or a network source.

With RTK, you also need either a GNSS base station or an RTK network service to get that correction data. Same with PPK. You're using either a base station or network service.

Third, when it comes to time in the field, PPK is the fastest setup, because you're handling the correction process after. It's a simpler workflow. RTK is what we'd call a medium setup. You need the correction source and to verify the connection between the drone and the correction source before and after the flight. Ground Control Points usually require the most setup time. You're placing visible targets around the site, measuring their exact location, and making sure they show up clearly in your images.

So accuracy isn’t one magic button. It’s a set of choices. Your equipment, your workflow, your environment, and your client’s requirements all play a role.

Next up, we’re going to get more concrete and more tactical. We’ll talk about Ground Sampling Distance, or GSD, and how image resolution relates directly to flight planning. This is where accuracy meets mission design, and it influences how you fly your patterns in the field.

I’ll see you in the next lesson.