Key takeaways
- They do not measure the same thing. A total station returns a small number of exactly located points. A drone returns millions of approximately located points describing a surface. Neither number is better; they answer different questions.
- The total station still wins wherever the drone cannot see: under canopy, inside structures, below grade, on vertical faces, and on any single point that has to be exact rather than representative.
- On most Alberta sites the two are not alternatives. Ground instrument work establishes and checks the control that makes the aerial surface trustworthy, and the drone covers the area the ground crew would have spent days walking.
- The cost shapes are opposite. Instrument work scales with the number of points; aerial work scales with area and barely at all with point count. That is the whole basis for choosing between them.
The question comes up on almost every first call with a survey or engineering buyer, and it is usually framed as a replacement decision. Do we still need the crew and the total station if we are flying the site?
The framing is the problem. A drone and a total station are not two ways of doing one job. They are two instruments that produce fundamentally different things, and a site that needs both is not a site that bought the wrong one.
What each instrument actually produces
A total station measures angles and a distance from a known setup to a specific target that a person deliberately placed. What comes back is a point, located to a precision far finer than anything aerial capture achieves, that means something because a human decided it meant something. The invert of that manhole. That property corner. The top of that footing. The instrument produces a small number of points that are exact and intentional.
Drone photogrammetry or LiDAR produces the opposite: an enormous number of points, none individually chosen, that together describe the shape of a surface. No single point in that cloud was picked by anybody, and no single point carries the confidence an instrument shot does. What the cloud has instead is completeness. It covers everything the sensor could see, including the parts nobody thought to shoot.
That is the entire distinction, and almost every sensible choice between the two falls out of it. If what you need is a handful of exact, meaningful points, the ground instrument is the right tool. If what you need is the shape of everything, the aircraft is.
Where the total station is still the right answer
Anywhere the sky cannot see
This is the big one and it is not close. A drone sees what is visible from above. Under a closed tree canopy, photogrammetry sees treetops and nothing else, and while LiDAR can push pulses through gaps in foliage it gets thinner and less reliable as the canopy closes. Inside a building, under a bridge deck, in a culvert, in a tunnel, beneath a roof, below water: none of that is aerial work. This is a hard limit of the method rather than a matter of equipment quality.
Single points that must be exact
Pipe inverts, structure offsets, bolt patterns, anchor positions, property corners and control monuments all share a property: they are individual points where being approximately right is being wrong. An aerial surface is an excellent description of where the ground generally is. It is not the tool for the one measurement a fabricator is going to cut steel against.
Vertical and overhanging surfaces
Aerial mapping is fundamentally good at surfaces that face upward. Retaining wall faces, structure sides, bridge soffits and undercut slopes are either poorly covered or invisible in a conventional nadir flight. Oblique aerial capture helps and is a real technique, but a ground instrument handles these directly with no argument about coverage.
When the aircraft cannot go up
Controlled airspace without authorisation, a site under a restriction, high wind, rain, or a small site where the mobilisation cost of aerial work outweighs the benefit. If four points are needed on a half-acre lot, a crew with an instrument is simply the faster answer.
Legal survey
Boundary and legal work has a defined process and a defined signatory, and aerial capture does not replace it. We covered that separation in who signs off drone survey data: a great deal of engineering and volumetric work needs no stamp at all, and boundary work needs an Alberta Land Surveyor regardless of what flew over it.
Where the drone wins, and wins decisively
Area
Point-by-point ground measurement scales linearly with the number of points, and describing a large irregular surface honestly needs a great many points. This is where a crew loses days. Aerial capture is close to indifferent to point count, which is why a site that would take a ground crew several days to model in reasonable detail is often a single flight.
Volumes and irregular surfaces
Stockpiles are the clearest case. A pile measured by instrument is a set of shots with interpolation between them, and the interpolation is where the argument starts. An aerial surface describes the actual pile, including the lumps and slumps nobody would have thought to shoot. See drone stockpile measurement for how that is handled against a defined base surface.
Anywhere a person should not be standing
Active haul roads, high walls, steep or unstable slopes, live process areas, contaminated ground. The safety argument is not marketing. A measurement taken from 60 metres up has nobody exposed to make it.
The visual record nobody asked for
A total station returns coordinates. A flight returns coordinates plus a dated, georeferenced orthomosaic of the whole site, which turns out to answer questions six months later that nobody thought to ask on the day. That by-product has settled more disputes than most of the deliberate deliverables.
The third instrument in the conversation
Most comparisons quietly leave out the GNSS rover, which is doing a lot of the work the total station gets credited with. A rover collects accurate discrete points far faster than an instrument, needs no line of sight to a setup, and is limited by needing usable satellite reception. Under canopy or against a tall structure it struggles where the total station keeps working. In open ground it is often the fastest way to collect the control and check points an aerial survey is built on. The realistic comparison on a modern site is a three-way one, not a two-way one.
How they are actually combined on one site
The productive arrangement is not either-or, and it looks about the same on most jobs.
- Ground instrument work establishes the control the flight is tied to, and just as importantly the independent check points the surface is verified against afterward. Without those, the aerial accuracy claim is unverified.
- The drone covers the area: topography, surfaces, volumes, imagery, the general shape of everything.
- The ground crew returns for what the aircraft could not see or could not resolve: inverts, under-canopy ground, structure detail, the points that have to be exact.
- Both sets land in one coordinate system and one vertical datum, which is where mismatches usually appear if the projection was never agreed. Our Alberta coordinate system picker exists for exactly that conversation.
Done this way the ground crew stops spending days walking a surface that a flight covers in one, and spends its time on the measurements only it can make. That is usually where the saving actually comes from, and it is not the saving people expect when they ask the replacement question.
A note on comparing accuracy numbers
Accuracy figures for the two methods are frequently compared as though they mean the same thing, and they do not. An instrument specification describes the precision of one measurement to one target. An aerial accuracy figure, such as the centimetre-grade results we describe in photogrammetry for civil engineers, describes how well a whole surface agrees with independent check points across a site. A method can be excellent at one and irrelevant to the other. The useful question is never which instrument has the smaller number, it is whether the deliverable meets the tolerance the decision actually requires, and how that was demonstrated.
Writing the scope
Most disputes on mixed jobs come from the scope never saying which method produced which number. Worth stating explicitly:
- Which features are captured aerially and which are ground-measured, named individually rather than implied.
- The tolerance required for each, since a topographic surface and a structure offset do not need the same one.
- The coordinate reference and vertical datum for everything, ground and air alike.
- Who provides the control and the independent check points, and that the check residuals are reported in the deliverable.
- What is explicitly excluded, particularly under-canopy ground, below-water bathymetry, interiors and anything below grade.
A scope that answers those five leaves very little room for the argument that starts with someone measuring a pipe invert off an orthomosaic.