A flight plan answers three questions before you take off: how far apart the lines go, how many photographs that produces, and whether the battery covers it. All three come out of the same four inputs — height, camera, overlap and site size — and the arithmetic fits on one page. The part people get wrong is not the arithmetic but the direction of the overlap: side overlap governs the spacing between lines, forward overlap governs the spacing along a line, and swapping them gives a plan that looks plausible and flies badly.
Overlap is a fraction of the ground a single photograph covers, so the footprint comes first. With
height H, sensor width Sw and focal length f:
How to get those camera numbers, and why the 35 mm equivalent focal length breaks the formula, is covered in how to calculate drone GSD. For the worked example below the camera is a 4/3 sensor 13.2 × 8.8 mm with a 12.29 mm lens, flown at 120 m. That gives a footprint of 128.9 m across and 85.9 m along.
At 70% side overlap, adjacent lines are 30% of the footprint apart, so
128.9 × 0.30 = 38.7 m between lines. At 70% forward overlap the aircraft advances
85.9 × 0.30 = 25.8 m between shots. Note what the percentage means: 70% overlap means
each frame shares 70% of its ground with the previous one, which is why the new ground per
step is the remaining 30%.
This is where the swap happens. It is tempting to apply forward overlap to the line spacing, because "forward" sounds like the along-track direction and lines are laid along-track. But line spacing is measured across the flight direction, between parallel passes, and that is the side overlap's job. Applying 70% forward overlap to a 128.9 m dimension gives the same number here only because both overlaps are 70%; set them to 80% and 60% and the two plans diverge immediately.
"Site width" is the dimension across the flight lines and "site length" is along them — the aircraft
direction, not the compass direction. Take a site 500 m wide and 400 m long:
⌈500 ÷ 38.7⌉ = 13 lines and ⌈400 ÷ 25.8⌉ = 16 shots per line, so
208 photographs. The ceiling function matters: 12.9 lines is 13 lines, never 12, or the last
10 m of the site is never photographed. Some planners add a line at each edge for insurance, and
on a site where the boundary is the deliverable that is the right instinct.
Check the coverage before believing the count: 208 frames × 128.9 × 85.9 m gross, times (0.3 × 0.3) net, is about 207 000 m² against a 200 000 m² site. The margin is thin, which is expected — a correct plan covers the site once with almost nothing to spare, because overlap is the redundancy.
Each pass is 400 m, so 13 passes are 5 200 m; the aircraft also has to cross between adjacent lines, 12 times at 38.7 m, which is another 464 m. Total 5 664 m. Forgetting the turn legs understates the distance by 8% on this plan and by far more on a long, narrow site where the turns dominate.
Divide by ground speed for flight time — but only by ground speed. A drone flying into a 5 m/s headwind at 12 m/s airspeed is making 7 m/s over the ground and needs proportionally longer. Compass heading does not change that; only ground speed does. At 29 km/h (8.06 m/s) the 5 664 m take 11.7 minutes of flying.
| Scenario | Line spacing | Shot spacing | Lines × per line | Photos | Distance | Time |
|---|---|---|---|---|---|---|
| 120 m, 70/70 | 38.7 m | 25.8 m | 13 × 16 | 208 | 5 664 m | 11.7 min |
| 120 m, 80/80 | 25.8 m | 17.2 m | 20 × 24 | 480 | 8 490 m | 17.6 min |
| 120 m, 60/60 | 51.6 m | 34.4 m | 10 × 12 | 120 | 4 464 m | 9.2 min |
| 60 m, 70/70 | 19.3 m | 12.9 m | 26 × 32 | 832 | 10 883 m | 22.5 min |
| 1 km × 1 km, 70/70 | 38.7 m | 25.8 m | 26 × 39 | 1 014 | 26 967 m | 55.8 min |
Read that table downwards and the economics become obvious. Doubling overlap from 60% to 80% multiplies the photo count by four and the flight time by two. Halving the height multiplies the photo count by four. The last row is the one that decides project feasibility: a square kilometre at 1 km² takes nearly an hour of flying, which is two or three batteries and a partial-return cycle in the middle.
Battery and return. Useful endurance is not the advertised flight time; it is the advertised time minus the reserve you fly to, and minus the climb. A plan needing 22 minutes on a nominal 25-minute aircraft does not fit, and a plan needing 56 minutes must be split into missions with enough overlap between them that the strips stitch together.
Turn and stabilise time. Between lines the aircraft must slow, turn and re-accelerate. With 2 seconds per turn the 12 turns above add 24 seconds — negligible here, but on a 60-line plan it is two minutes and worth allowing for.
Terrain. A constant height above the take-off point is not a constant height above the ground. Over a hillside the footprint shrinks on the uphill side and grows downhill, so both the coverage and the GSD drift. Terrain following fixes it; nothing in the arithmetic does.
Wind and geometry. Lines laid at a bearing that crosses the prevailing wind give the smoothest images, but lines square to a straight site boundary give the fewest of them. On an irregular site, plan the rectangle that contains it and accept the photographs of the fields next door — masking during processing is cheaper than flying a polygon.
Sixty percent overlap covers the ground; it does not survive the stitch. The 60/60 row above is a valid arithmetic answer and a common marketing figure, but 60% forward overlap gives a photogrammetric matcher little to work with on sparse, uniform terrain — a pasture, a beach, fresh asphalt. For engineering work, 75% to 80% forward with 65% to 70% side is the range that survives a change of season and a tired processor. Plan at 70/70, spend the extra photos.
Fix the GSD first, because it sets the height. Fix the overlap second, from the deliverable — an orthophoto tolerates less overlap than a dense point cloud. Compute spacing, then counts, then distance, then time. Then compare the time with endurance and split the plan if it fails. Doing it in that order means every number you compute afterwards is still valid; doing it in the reverse order means discovering the plan does not fit after you have already optimised it.
CoordCalc publishes free surveying and coordinate calculators. Results are for reference — confirm against the governing standard before construction or procurement. About · Privacy & cookies · Terms of use · Developers