A mapping flight isn't taking pictures of a site. It's taking pictures of the same parts of a site from several different positions, so that software can work out the three-dimensional shape from the parallax between them.
Overlap is how much of each photo repeats in the next one. It comes in two directions:
- Front overlap (forward, or end lap) — between consecutive photos along the same flight line. Controlled by how often the camera triggers versus how fast you fly.
- Side overlap (side lap) — between adjacent parallel flight lines. Controlled by how far apart the lines are spaced.
Every point on the ground needs to appear in enough photos, from sufficiently different angles, for the software to triangulate it. Where that fails, you don't get a slightly worse model — you get holes, warping, and a mosaic that visibly tears.
Starting numbers
These are starting points rather than rules. Most of them follow Pix4D's published image-acquisition guidance; check your own processing software's recommendations, since they differ. Push the numbers up when the ground is difficult, not when you want a nicer result on easy ground.
| Scenario | Front | Side |
|---|---|---|
| General orthomosaic, textured ground | 75% | 60% |
| Agricultural fields | 80% | 80% |
| Forest and dense vegetation | 85% | 85% |
| Uniform surfaces (sand, snow) | 85% | 70% |
| Vertical structures, facades | 85%+ | 80%+ |
The pattern behind the table: overlap requirements are driven by how much visual texture the ground has, not by how important the job is.
Why featureless ground needs more
Photogrammetry works by matching identifiable features between images. A gravel yard is full of them — every stone is a distinguishable landmark. Fresh snow, calm water, a uniform sand flat, or a newly poured slab have almost none. Two photos of blank white look identical no matter where they were taken from, so the software cannot establish how they relate.
More overlap is the partial fix: it gives the matcher more chances to find the few features that do exist, and shortens the distance it has to bridge between them. It is a mitigation, not a cure. Some surfaces simply do not reconstruct well from imagery, and no overlap setting changes that.
Moving subjects cause a related failure. Tree canopy in wind, traffic, livestock, and water surfaces all move between exposures, so the same feature appears in different places in different photos. High overlap helps the software identify and reject those points as inconsistent.
What high overlap actually costs
It is tempting to set everything to 90% and stop thinking about it. The cost is real and it compounds in both directions at once.
Front overlap costs photos along the line: going from 75% to 85% cuts the distance between exposures from 25% of the footprint to 15%, which is roughly two-thirds more shots per line. Side overlap costs flight lines: going from 60% to 80% cuts line spacing from 40% of the footprint width to 20%, which doubles the number of lines needed to cover the same area.
Do both and you have roughly tripled the image count for the same site. That means:
- More battery swaps, or a site you can no longer cover in the window you have.
- Processing time that scales worse than linearly — doubling images can more than double the hours.
- Storage and transfer that stop being trivial.
- A higher chance of the camera not keeping up with the trigger interval, which produces its own gaps.
That last one is worth checking on any fast, high-overlap plan. If the required trigger interval is shorter than the camera's actual write cycle, the aircraft flies past exposures it never took. The plan looks correct; the dataset has holes in it. Slowing the flight is the fix.
Practical habits
Fly past the boundary. Coverage degrades at the edges of a survey, where points appear in fewer images. Extend your area beyond what you actually need — a couple of flight lines of margin — so the usable region sits in the well-covered middle.
Add a crossing pattern for 3D. For structures and volumetrics, flying a second grid perpendicular to the first gives every surface a much better range of viewing angles. This helps more than pushing overlap higher on a single-direction grid.
Keep overlap honest on slopes. Overlap percentages assume a footprint size, and footprint size depends on height above ground. On terrain that rises toward the aircraft, the footprint shrinks and your real overlap drops below what you planned — the same effect that makes terrain follow matter.
Check coverage before you leave site. Overlap failures are cheap to fix with a re-fly and expensive to fix from the office. If the site is far away or access is restricted, it's worth confirming before packing up.
DronePoint derives flight-line spacing and photo intervals from the overlap percentages you set, using your drone's real sensor footprint, so the spacing matches what you asked for rather than a generic assumption.
Related: ground sample distance sets your resolution ceiling — the other number worth deciding before takeoff.
Follow DronePoint
See more of these notes in your Google Search results.
Add DronePoint as a preferred source