Working in drone surveying and photogrammetry: the trade

Of all the career paths open to a drone pilot, surveying is the one that looks least like what people picture when they think about «working with drones». There are no spectacular shots and no video. There is a construction site at seven in the morning, a flight plan loaded the night before, targets spread across the ground and, afterwards, four hours in front of a computer processing what took twenty minutes to fly.

It is also the one with the healthiest relationship between effort and job: the client is not hiring a flight, they are hiring data to a given accuracy, and they will need that data again next month and the month after.

This article describes what the trade is really like —what you deliver, who pays, what equipment you handle and what training is required— for anyone weighing up whether it fits them. If what you want is to understand the technique itself, we already have what drone photogrammetry is; this is the other half, the work.

What actually gets delivered on a surveying job

The most common confusion for a beginner is believing the deliverable is the photos. It is not. Depending on the job, a surveying client expects to receive some combination of this:

• Georeferenced orthophoto: the image of the ground corrected geometrically, so that measurements can be taken on it.
• Point cloud: millions of points with X, Y and Z coordinates. It is the raw material for almost everything else.
• Digital terrain or surface model: the representation of the relief, with or without vegetation and buildings.
• Contour lines and drawings in whatever format the client uses, usually CAD.
• Volume calculations: how much material is in a stockpile, how much has been extracted from a quarry this month, how much is left to move in an excavation.

That last point is what best explains why the work repeats. A quarry needs its extracted volume every month. A construction site needs to certify progress. The drone turns a measurement that used to cost two days of a surveyor into one morning, and that is why it gets hired on a recurring basis rather than once.

Accuracy does not come from the drone, it comes from the ground control points

This is the detail that separates someone selling «flights» from someone selling surveying. A photogrammetric flight without ground control gives you a pretty model that lies geometrically. Accuracy comes from ground control points measured with GNSS spread across the area, or from RTK/PPK systems that correct the drone’s own position during the flight.

Whoever is hiring knows this perfectly well. The moment they ask «what accuracy do you guarantee me?» and the answer is vague, the conversation is over. It is by far the part of the training that pays off most for anyone doing this for a living.

Photogrammetry or LiDAR

Photogrammetry reconstructs the ground from overlapping photos: it is cheaper, more accessible and enough for most jobs. LiDAR emits laser pulses and measures the return, which lets it get through vegetation and reach the ground where photogrammetry only sees treetops. That is why it rules in wooded areas, infrastructure corridors and forestry mapping.

It is not a matter of taste, it is a matter of terrain.

What a job looks like from start to finish

It is worth seeing the whole thing, because the flight is the short part:

1. Define the target accuracy. You do not fly the same way for a preliminary study as for certifying construction progress. That determines the GSD —the pixel size on the ground— and from there the flight altitude.
2. Study the area before going. Airspace, obstacles, access and permits if there is private property or the environment is controlled.
3. Plan the flight. Sufficient front and side overlap, the pattern of passes, and cross or oblique flights when there are facades or vertical slopes.
4. Place and measure the ground control points. Spread out, not clustered, keeping some back as check points to verify against afterwards.
5. Fly. Usually the quickest part of the day.
6. Process. Orientation, densification, cloud generation, classification and filtering.
7. Check the quality. Compare the model against the check points that were not used to orient it. This step is what turns a file into a defensible deliverable.
8. Deliver in the client’s format, with a report stating what accuracy was achieved and how it was verified.

If any of those steps is missing, the work still looks good on screen and is useless for what the client needed it for. That is, in practice, why jobs get redone and clients get lost.

The four things that ruin a survey

None of them has anything to do with the drone, which is why they catch out anyone coming from hobby flying:

• Surfaces without texture. Water, uniform sand, snow, freshly laid asphalt. Photogrammetry needs to find the same point in several photos, and there are no points to find there.
• Hard shadows and changing light. Flying at midday with intermittent cloud leaves a mosaic with patches that do not match.
• Wind and moving vegetation. Treetops move between one pass and the next and the algorithm reconstructs them badly.
• Badly distributed control points. All bunched in the middle of the plot gives you a model that bends at the edges.

Knowing how to read the conditions and decide that today you are not flying costs you a trip; flying anyway costs you the whole job and the client’s confidence.

Who hires this in Spain

Worth looking at, because it shapes the kind of professional life you are going to have:

• Engineering firms and surveying practices. The most frequent client. They tend to subcontract the flight and keep the processing, or the other way round, depending on what they lack.
• Construction companies. Site monitoring and progress certification. Continuous work for as long as the project lasts.
• Quarries, mining and aggregates. Periodic measurement of stockpiles and faces. The recurring job par excellence.
• Public administrations and councils. Mapping, urban planning, inventories. Long cycles and tender documents, but volume.
• Agriculture and forestry. When the survey crosses over with vegetation indices, the job starts to look more like precision agriculture.

The difference with film work is obvious: here the client is technical, knows what they are asking for and values repeatability over creativity. If you prefer the other profile, see what we cover in career paths for drone pilots.

What the legal side will require of you

Drone surveying rarely fits in the open category. As soon as the area to be surveyed is large, you have to fly beyond what comfortable visual line of sight allows, or the ground is in a populated environment, or you have to overfly a site with people working. All of that pushes you into specific category.

The usual route is the operational declaration under a standard scenario. And here it is worth bearing in mind this year’s change: since 1 January 2026 the national STS-ES scenarios are no longer valid. You have to declare under STS-01 —flight within visual line of sight over a controlled area in a populated environment, with a C5 class drone— or STS-02 —beyond visual line of sight, with observers, in a sparsely populated environment and C6 class—. Either one requires the STS theoretical knowledge certificate issued by AESA.

Add the cross-cutting parts: registration as a UAS operator if you are going to invoice on your own account, civil liability insurance and, when the area is near airports or in controlled airspace, prior coordination. That last procedure is the one that delays the most projects.

The equipment, and why you do not buy it first

An expensive and frequent mistake: buying the drone before knowing which scenario you are going to declare under. The aircraft’s class marking —C5, C6— is not a manufacturer’s ornament, it is what makes your operational declaration possible or impossible. And if your specialisation is going to be surveying, the RTK/PPK module and the quality of the GNSS receiver weigh far more on the final result than the camera’s megapixels.

First you decide the type of work, then the scenario, then the aircraft. In that order.

What training the sector asks for

There are three blocks and none is optional if you want to get paid for this.

The aeronautical base. Open category for the minimum permission and, above all, the STS certificate so you can work in specific category. Without it, half the real jobs are legally out of reach. It is covered in the professional drone pilot course and in the STS certifications course.

Field technique. Flight planning with the right overlap, placing and measuring ground control points, managing flight altitude according to the GSD the client demands, and controlling the conditions that ruin a survey —wind, hard shadows, water or textureless sand—.

Processing. This is where most of the time goes and where it shows who knows their trade. Point cloud generation, classification, vegetation filtering, model generation, quality control against the control points and export to whatever format the client uses. The photogrammetry and surveying course is built on those three blocks, with real flight practice.

And a warning about the market

This specialisation has a higher barrier to entry than film work, and that is good news for whoever crosses it. The client tells perfectly well the difference between someone who delivers a point cloud with an accuracy report and someone who sends a file with no quality control. In the first case they come back; in the second, they do not.

The practical consequence is that the training effort pays back sooner here than in almost any other path: you are not competing with hundreds of hobbyists with a drone, you are competing with very few people who can deliver defensible data.

If you are weighing up whether this is your route or one of the others, in which drone course you need depending on what you want to do all the specialisations are compared. And if you have already decided, tell us where you are starting from —surveying, construction or from scratch— and we will tell you which part of the route you can skip.