Products · VTOL Fixed-Wing

Orthophoto Acquisition Configuration

A correct orthophoto is specified, not processed into existence. Overlap, GSD, control and datum are fixed before the flight or they cannot be recovered.

7 of 10Method note

Overview

State the target GSD and the required horizontal and vertical accuracy; both drive altitude and control density.
Nadir-only capture limits facade detail; if buildings matter, add an oblique pass rather than stretching the ortho.
Homogeneous surfaces (water, fresh tarmac, uniform vegetation) will not tie without enough texture or GCPs.
Confirm the tile scheme and naming convention with whoever consumes the data.

An orthophoto is a product, not a picture

The difference between a set of aerial images and an orthophoto is geometric correction: each pixel is placed where it would fall on a flat reference plane, removing the displacement caused by terrain relief, sensor tilt and perspective. That correction is what makes the output measurable, and it is why an orthophoto is specified by its ground sampling distance and its positional accuracy rather than by how it looks.

This has a direct consequence for planning. Because the product is geometric, the acquisition must be planned to support the reconstruction that produces it — sufficient overlap, sufficient positional knowledge, and enough images over each point that the solution is over-determined rather than barely determined. A flight that produces attractive images can still fail to produce a usable orthophoto, and it usually fails in the places that matter most: building edges, tall structures and steep terrain.

So specify the output first. Ground sampling distance, horizontal and vertical accuracy, and the coordinate system and datum the deliverable must be in. Those four numbers determine the sensor, the altitude, the overlap, the georeferencing approach and the processing chain. Everything else is implementation detail.

Ground sampling distance and the altitude that delivers it

Ground sampling distance is the distance between pixel centres measured on the ground, and it is set by sensor resolution, focal length and flight altitude. It is the single most quoted orthophoto specification and the one most often specified without checking what it costs. Halving ground sampling distance quadruples the number of pixels per unit area and roughly quadruples the number of flight lines needed to cover the same ground — which quadruples flight time.

Specify the coarsest ground sampling distance that still resolves the smallest feature the deliverable must support. For site-level mapping, that is often much coarser than teams assume. Reserve fine ground sampling distance for the subset of the area where it is actually needed — a two-tier acquisition over the same site is usually far more economical than a uniform fine pass everywhere.

Note also that ground sampling distance is quoted at a nominal altitude over flat ground. Over sloping terrain, actual ground sampling distance varies across the strip, and over tall structures it varies dramatically. Where the deliverable concerns building footprints or roof detail, that variation — and the relief displacement that comes with it — is precisely why oblique imagery is collected alongside the nadir pass.

Overlap, georeferencing and the accuracy budget

Overlap is what makes reconstruction possible: each point must appear in enough images, from enough different viewpoints, for the geometry to be solved. Standard practice uses substantial forward overlap along a line and meaningful side overlap between lines. Reducing overlap to save flight time is the most common cause of reconstruction gaps, and those gaps appear in the deliverable as holes or as locally degraded accuracy.

The trade is real but asymmetric. Overlap above the necessary level costs flight time linearly and improves reliability with diminishing returns. Overlap below it does not degrade gracefully — it produces failures. Set it at a level appropriate to the terrain: flat, uniform ground tolerates less; terrain with relief, water or repetitive texture needs more, because those are exactly the surfaces where matching is unreliable.

Georeferencing deserves an explicit accuracy budget. Positional accuracy in the final product comes from the combination of the aircraft's GNSS solution, the georeferencing method, and any ground control. Direct georeferencing reduces field work but its accuracy depends on the onboard solution and its calibration. Ground control anchors accuracy in a verifiable way but costs field time. Most programmes use both, and the mix should be decided against the stated accuracy requirement rather than by habit. Our custom data acquisition standards work covers setting these parameters as a documented specification.

Conditions that quietly degrade the result

Motion blur is the most common silent killer of orthophoto quality. It arises when the aircraft moves too far during an exposure, and it is a function of ground speed, exposure time and shutter efficiency. Faster cruising platforms need shorter exposures; low light forces longer ones. The two meet at a limit, and beyond that limit the images look acceptable on screen and reconstruct badly.

Sun angle and shadow length matter more than most specifications acknowledge. Long shadows hide detail, reduce matching reliability and make interpretation harder. High sun reduces shadows but can flatten texture. For most mapping work, a mid-morning to mid-afternoon window with the sun reasonably high is the practical choice, and consistency of sun angle across epochs matters if the outputs will be compared.

Surface conditions matter too. Water, glass, fresh asphalt and repetitive texture are all poor matching surfaces, and they produce local reconstruction failures. Where a site contains significant areas of these, expect higher overlap requirements and plan for local gaps rather than assuming a uniform result.

Configuring an orthophoto programme

The configuration we would propose starts from the four output specifications and works backwards: sensor and altitude from ground sampling distance, overlap from terrain and surface conditions, georeferencing from the accuracy budget, and sortie plan from area and wind. Where the site has both open ground and detailed structures, a two-tier or nadir-plus-oblique design usually beats a uniform approach on cost.

Verification should be built into the plan rather than performed at the end. Check coverage and image quality per sortie, verify reconstruction on a subset before flying the whole block, and hold a defined accuracy check against independent check points before delivery. Our test and verification methods note describes how we structure that.

Send your area, required ground sampling distance, accuracy target, coordinate system and terrain profile to our engineering team. We will propose the acquisition design and the sortie estimate, and we will flag which parts of your site will need additional overlap or a second pass.