Drone Mapping & Flight Time

Photo count scales as 1/altitude squared, and the shutter — not the battery — usually sets your flight time on a detailed survey.

m above ground
hectares
cm/px, to solve for height
%
%
seconds between shots
%

Altitude against everything else

What overlap actually costs

Where the camera becomes the limit

Photo count scales as 1/altitude², which governs everything else

The footprint grows in both dimensions with height, so the area each photo covers grows as the square. A 10 hectare survey at 30 m needs exactly 16× the photos of the same survey at 120 — and (120/30)² is 16. The corollary is the one people feel: halving the ground sample distance quadruples the photos, the flight time and the processing.

And overlap is brutally non-linear, because the multiplier is 1/((1−front)(1−side)). Going from 75/65 to 85/75 sounds like a ten-point nudge and is 2.33× the photographs. Front overlap alone from 75 to 85 isn't "13% more" — it's 67% more, because the cost is one over the remaining fraction, and that fraction is what shrinks.

The constraint people don't expect is the camera, not the battery — and it binds far more often than I first assumed. Photos must be spaced by the footprint × (1−front overlap), and the shutter fires only so often, so ground speed is capped at spacing ÷ interval. On a consumer Mavic 3 at a 2 s interval and 75% front overlap, that ceiling stays below the aircraft's own 15 m/s until about 113 m — essentially the entire legal band under a 120 m limit. At 20 m with 85% overlap it's 1.6 m/s.

So for most real missions the shutter sets the flight time, not the battery and not the airframe. A faster capture interval is worth more than endurance on a detailed survey — which is exactly what separates the enterprise bodies from the consumer ones, and it isn't the spec anyone compares.

  • GSD is not accuracy, and conflating them is the commonest mapping mistake. GSD is how much ground one pixel covers; positional accuracy is where that pixel actually is, set by the aircraft's positioning and ground control. A 2 cm GSD without RTK can still be metres out in absolute terms.
  • Raise side overlap before front overlap on difficult terrain. Front overlap costs shutter interval; side overlap adds whole flight lines. But side overlap is what keeps reconstruction stable across a strip, so over trees or broken ground 65% is a floor rather than a target.
  • Fly in even light, not bright light. Hard shadows move between lines and confuse feature matching — a flat overcast day beats a bright afternoon at the same settings, and it does more for the result than 75 versus 85% overlap.
  • Battery count follows the same inverse square, so the altitude decision is really a decision about how many times you land.

How to use

  1. Pick the altitude from the resolution you need, not from habit.
  2. Check whether the capture interval caps your ground speed.
  3. Raise side overlap before front overlap on difficult terrain.
  4. Fly in even light rather than bright light.

Frequently asked questions

How does altitude affect the number of photos in a drone survey?

As one over the square. The footprint grows in both dimensions with height, so the area each photo covers grows as the square of the altitude — a ten hectare survey at thirty metres needs exactly sixteen times the photos of the same survey at a hundred and twenty. The corollary is what people feel in practice: halving the ground sample distance quadruples the photos, the flight time and the processing.

What is ground sample distance?

How much ground one pixel covers, usually quoted in centimetres per pixel. It is the sensor width times the altitude, divided by the focal length times the image width in pixels. A four-thirds sensor at 12.29 mm focal length gives about 2.68 centimetres per pixel at a hundred metres, which matches the figure published for that camera.

Is ground sample distance the same as accuracy?

No, and conflating them is the commonest mapping mistake. Ground sample distance is how much ground one pixel covers; positional accuracy is where that pixel actually is in the world, and it is set by the aircraft positioning and any ground control rather than by the camera. A two-centimetre GSD from a drone without RTK can still be metres out in absolute terms.

How much does increasing overlap cost?

Far more than the numbers suggest, because the multiplier is one over the product of the remaining fractions. Going from 75/65 to 85/75 sounds like a ten-point nudge and is 2.33 times the photographs. Front overlap alone from 75 to 85 per cent is not thirteen per cent more, it is sixty-seven — the cost is one over what is LEFT, and that is what shrinks.

What overlap should I use for drone mapping?

Seventy-five per cent front and sixty-five side is the usual starting point, and it is a reasonable default for open ground with texture. Raise it for uniform surfaces, vegetation, water or anything with few distinct features, and raise the SIDE overlap first — front overlap costs shutter interval while side overlap adds whole flight lines, but side overlap is what keeps the reconstruction stable across a strip.

Why is my drone flying so slowly on a mapping mission?

Almost certainly the capture interval rather than anything you set. Photos have to be spaced by the footprint times one minus the front overlap, and the camera can only fire so often, so ground speed is capped at that spacing divided by the interval. At twenty metres with eighty-five per cent front overlap and a two-second shutter, the ceiling is 1.6 metres per second.

Is the battery or the camera the real limit on a survey?

The camera, far more often than people assume. On a consumer Mavic 3 at a two-second interval and seventy-five per cent front overlap, the speed ceiling stays below the aircraft own fifteen metres per second until about a hundred and thirteen metres — which is essentially the entire legal altitude band under a hundred and twenty metre limit. The shutter sets the mission time on most detailed surveys.

Is a faster capture interval worth more than longer endurance?

On a detailed survey, yes. If the shutter is capping your ground speed then extra battery simply buys more time flying slowly, whereas a faster interval raises the ceiling and shortens the mission outright. It is the specification that separates enterprise bodies from consumer ones and it is rarely the one people compare, because endurance is the number printed on the box.

How many batteries will a drone survey need?

It follows the same inverse square as the photo count, so the altitude decision is really a decision about how many times you land. A ten hectare survey that takes one battery at a hundred metres can take several at thirty, for the same area and the same overlap — the only thing that changed is the resolution, and it changed the workload by the square.

What time of day is best for drone mapping?

Even light rather than bright light. Hard shadows move between flight lines and confuse the feature matching, so a flat overcast day produces a better model than a bright afternoon at identical settings. That single choice does more for the result than the difference between seventy-five and eighty-five per cent overlap, and unlike overlap it costs nothing in flight time.

Why does side overlap matter more than front overlap on rough ground?

Because it is what ties adjacent strips together. Front overlap gives you many views along a line, which is cheap and helps within a strip, but if neighbouring lines share too little the reconstruction can drift between them — and broken ground, trees and uniform surfaces are exactly where that drift appears. On difficult terrain sixty-five per cent side overlap is a floor rather than a target.

What altitude should I fly for a given resolution?

Work backwards from the ground sample distance: altitude equals the target GSD times the focal length times the image width, over the sensor width. Then check what that does to the photo count, because the relationship is quadratic — asking for twice the resolution does not cost twice the flying, it costs four times, and it may move you into the region where the shutter caps your speed as well.

🔒 This tool runs entirely in your browser. Nothing you enter is uploaded, logged, or stored.