2026 results

What the drone saw
in the vineyards.

A first season on working vineyards, not in a lab. Results are anonymised and presented by grape variety. We also show what did not work: that is how we improve.

9 + 2estates and cooperative wineries
29plots monitored
39 haflown over
109flights, from May to October

Downy mildew · Marsanne plot

Two hotspots found, with no clue given

The estate reported downy mildew without saying where. The drone had to find the areas on its own.

Aerial photo of the plot
01 · FROM THE SKYNothing stands out to the eye.Flight of 10 July, after a rainy spell.
NDRE map: two red areas
02 · WHAT THE DRONE FLAGS2 areas at −13%Foliage health index below the plot average.
Hotspots found on the ground within the flagged areas
03 · CONFIRMED ON THE GROUND2 hotspots, both inside the areasAt the next flight, the gap reached −22% and −28%, over 0.22 ha.

For a winegrower, that is the difference between treating two hotspots and treating the whole plot.

Downy and powdery mildew · 16 plots

The drone also tells you where not to go

On ten plots, no major problem was detected, which matched what the estates saw on the ground where they had inspected their vines. On the other six, the drone flagged precise areas. For a winegrower short of time in summer, knowing that a plot does not need a walk-through is worth as much as an alert.

10 / 16

plots with no major problem

< 1 ha

to inspect as a priority, out of about 21 ha flown over

Flavescence dorée · Marselan plot

Mapping where the disease is

29 rows, 1.3 ha, a heavily affected plot: 584 vines marked on the ground.

The Marselan plot seen from the sky at the end of September
01 · FROM THE SKYSome red shows here and there, but all Marselan vines redden a little in autumn.
Sections of row with the most reddening, flagged by the drone
02 · WHAT THE DRONE FLAGSEach row split into 3 m sections; in red, the most suspicious vines.
Density of vines marked on the ground during field scouting
03 · THE FIELD SURVEY6 of the 8 areas flagged before scouting had a higher density of affected vines than the plot average.

At this stage, the drone finds where the disease is concentrated, not yet every single vine.

The blind test

Twenty alerts locked before we knew which vines were marked

On a lightly affected Marselan plot, the main hotspot fell within a flagged area. On a neighbouring Grenache blanc plot, two isolated vines were missed. Too few to draw conclusions, enough to shape the next steps.

Map of empty vine positions on a Grenache blanc plot
Counting missing vines13%of vine positions empty on this Grenache blanc plot, about 800 vines to replant. A plant order prepared without counting on foot.
What the drone cannot see yet

An isolated vine in the early stages of the disease still looks mostly green from above: the drone does not flag it. A flight too early in the morning, with a low sun and long shadows, blurs the measurement. One season is not enough to solve all of this.

In 2027, we go further

Fly at the right time
Mid-September, with a high sun.

Fly twice
A diseased vine reddens before the others.

Better ground positioning
Centimetre GPS, not a phone.

Plots in between
20 to 50 affected vines, hotspots just starting.