C-DRONE GUIDE · 1 JUNE 2026
Drones in precision agriculture: mapping, nitrogen, spraying
Long a farm-show gadget, the drone has become a production tool: biomass maps for nitrogen modulation, emergence counts, wildlife damage assessment, and now spraying authorised on steep slopes. This guide sorts out what genuinely pays per hectare, what remains a comfort, and what French regulations allow — or not — in 2026.
What the sensor sees that the eye cannot
The heart of drone-based precision agriculture is the multispectral sensor: beyond visible light, it records near-infrared, which vegetation reflects all the more strongly as it is active. The computed indices — NDVI and above all NDRE, more discriminating on dense canopies — translate this reflectance into vigour maps at a few centimetres' resolution, where free satellite imagery (Sentinel-2) tops out at 10 m and depends on cloud cover. On a 20 ha field, a 25-minute flight produces a map where every heterogeneity becomes visible: soil crusting, failed drilling, an incipient pest attack, a leaking drainage line.
The map, however, is only a decision aid: a low NDVI says "there is a problem here", never which one. The value comes from crossing it with agronomy — field history, soil type, crop stage — and from the targeted field walk the map enables: instead of pacing 20 ha, the farmer goes straight to the three abnormal zones. That is the right use of drones in agriculture: not replacing the agronomist, but directing their gaze.
The uses that pay: nitrogen, counts, assessments
Three uses have proven their profitability in France. Nitrogen modulation first: a late-winter biomass map on oilseed rape adjusts the nitrogen dose zone by zone — measured above-ground biomass converts directly into nitrogen already absorbed — and wheat maps at key stages feed decision-support tools. Observed savings: 20 to 40 nitrogen units per hectare in good years, i.e. €25 to €50/ha at fertiliser prices, for a service billed €8 to €15/ha. The prescription file is exported in standard formats (ISOXML, shapefile) directly readable by recent spreader and sprayer terminals.
Second profitable use: density counts. Beet, maize or sunflower emergence counted by AI on high-resolution imagery, to decide on re-drilling in days rather than weeks — and to document the decision for the insurer or seed company. Third, fast-growing use: damage assessment — wildlife, hail, flooding. A georeferenced orthophoto measures destroyed areas to the square metre, an objectivity that speeds up compensation from hunting federations and insurers; several agricultural loss adjusters have bought their own equipment. Then come the comfort uses: irrigation monitoring, grassland inventories, trial-plot monitoring for seed companies and technical institutes.
Drone spraying: what France allows in 2026
It is the most asked — and most regulated — question. The principle set by the French rural code remains a ban on aerial spraying of plant protection products. But the framework has evolved: after a three-year national experiment evaluated by the ANSES food-safety agency, law 2025-365 of 23 April 2025, completed by its spring 2026 implementing texts, opened a permanent derogation for drone spraying on plots with slopes of 20% or more, banana plantations and grapevine mother stock, limited to biocontrol products, products authorised in organic farming and low-risk products. Steep vineyards — Alsace, Côte-Rôtie, Beaujolais, Banyuls — and banana plantations are the first concerned: the drone replaces the backpack sprayer, with an obvious safety gain for operators who no longer scramble, tank on back, across slopes where tractors overturn.
In practice, the operation requires a certified agricultural drone (DJI Agras T50/T70 dominate, €25,000 to €45,000 equipped), a specific-category authorisation, the operator's Certiphyto spraying licence, and compliance with each authorisation's conditions (buffer zones, weather, record-keeping). Spreading solid non-phytosanitary products — cover-crop seeding, trichogramma capsules against corn borer, targeted fertiliser — is much more widely practicable and is a real market: drone seeding of cover crops before harvest is growing fast, billed at €25 to €45/ha.
Equipment and providers: the right 2026 choices
For mapping, the de facto standard is the integrated multispectral drone, DJI Mavic 3M class (4 spectral bands plus RGB, sunlight sensor, optional RTK module, around €4,500 to €6,000 excl. VAT): it covers 100 to 200 ha per flying day. Above that, fixed wings (senseFly, Delair) keep the advantage on very large areas. Processing goes through dedicated platforms (Solvi, Pix4Dfields, DJI Terra or the integrated services of farm-supply distributors) which generate indices and prescription maps without photogrammetry expertise.
The real question for a farmer is buying versus hiring. Buying makes sense from 300 to 400 ha worked in several annual passes, or through a machinery cooperative: budget €6,000 to €10,000 of investment, plus pilot training (the A2 certificate covers most rural fields, subject to the 120 m maximum height and the zones of the official drone map — mind the very-low-altitude military corridors that criss-cross the French countryside). Hiring costs €8 to €15 per mapped hectare with a call-out minimum, and €25 to €45/ha for spreading: most farms come out ahead without tying up capital or training time. Cooperatives and merchants are moreover structuring "drone included" offers within their agronomy advisory contracts.
Profitability: the honest per-hectare maths
Let us run the numbers on a typical case: 150 ha of arable crops, two annual mapping passes (oilseed rape at winter's end, wheat at stem extension) hired at €10/ha, i.e. €3,000 per year. Against that: the modulated-nitrogen saving (€20 to €40/ha on heterogeneous fields, i.e. €1,500 to €4,000 on the half of the acreage concerned), the yield gain from zones corrected in time, and the diffuse savings — a re-drilling avoided, a damage claim better compensated. In favourable years, the return is two to four times the outlay; in years when the fields are uniformly fine, the map brings almost nothing. Agricultural drone profitability is statistical, not systematic — whoever promises otherwise is selling dreams.
Three conditions maximise the odds: genuinely heterogeneous fields (varied soils, contrasting yield history — combine yield maps tell you), spreading equipment able to modulate (compatible terminal, variable-rate control), and the discipline to act on the maps within days of the flight, because a biomass map has an agronomic shelf life of one to two weeks. Without those three conditions, start small: a damage assessment or an emergence count when the occasion arises, to judge on evidence with a local provider before scaling up.