C-DRONE GUIDE · 7 SEPTEMBER 2026
Sugarcane seen from a drone: vigour, ratoons and biomass in Réunion and the French Antilles
In Réunion, sugarcane covered 21,393 hectares at the 2020 agricultural census, or 55% of the department's utilised agricultural area, farmed by 2,718 holdings, more than 70% of them under eight hectares. In Guadeloupe, 3,087 holdings across 12,417 hectares. In Martinique, cane remains the second crop behind bananas. Yet these are among the hardest fields to inspect in all of French agriculture: mature cane stands 2 to 5 metres tall, which makes it impossible to read the inside of a block from its edges, and the plots are fragmented, steep, often perched up to 800 metres above sea level on Réunion's slopes. After three disastrous campaigns — 1.139 million tonnes in Réunion in 2024, 30% below the ten-year average, then cyclone Garance in 2025 — knowing exactly where a field is failing is no longer a luxury. Here is what a drone survey shows, and what it will never show.
Published on 7 September 2026, reviewed on 7 September 2026 — regulations in force as of September 2026.
A sector under pressure, on fields nobody can actually see
Public figures from the three cane-growing departments tell the same story. In Réunion, output fell to 1.139 million tonnes in 2024, 30% below the ten-year average, at a sugar content of 13.0% (Insee, 2024 economic review). In Guadeloupe, the 2024 campaign ended with 264,000 tonnes of cane milled and 14,600 tonnes of sugar, the worst result ever recorded. In Martinique, 206,431 tonnes milled, sucrose content down to 10.1% against 12.1% the year before — its lowest in a decade — 660 tonnes of sugar and 91,600 hectolitres of pure alcohol of rum, after an early-year drought followed by hurricane Beryl in late June.
Weather is not the whole story, but it hits hard: cyclone Garance, on 28 February 2025, with winds above 200 km/h, caused what Réunion's chamber of agriculture put at 151.6 million euros of farm losses, some 80 million of them in the cane sector alone, which accounts for 52.9% of the island's farmed area. In that context every tonne counts and every unproductive hectare is paid for. Yet cane is one of the few major crops a field technician simply cannot walk: at 3 metres tall, in a dense stand, you literally see nothing beyond the first row. Inspections happen at the edge, at the end of a row, or by climbing onto a loader bucket. The middle of a block stays a blind spot — and that is often exactly where things go wrong. The rules for flying in the overseas departments are the same as in mainland France, with terrain specifics of their own: our guide to drone services in France's overseas departments covers that point.
What the flight reveals: variability, ratoons, gaps, lodging
A mapping flight produces two complementary deliverables: a high-resolution orthophoto, in which every row is visible, and a vegetation-index map from a multispectral sensor, which turns canopy vigour into a colour gradient. On a cane block, that feeds four concrete decisions.
- Within-field variability: spotting the zones that close the canopy poorly, often tied to soil depth, waterlogging or a fertilisation shortfall, so a nitrogen top-up can be targeted instead of treating the whole block.
- Ratoon quality after cutting: cane regrows from the stool, and the vigour of those ratoons drives the following campaign. A sluggish regrowth in one sector, invisible from the road, is obvious on an index map a few weeks after the cut.
- Gap counting: holes in the row — dead stools, harvester damage, compaction — can be counted on the image and converted into a percentage of bare soil, the key figure behind the replanting decision.
- Lodging after a cyclone or heavy rain: flattened cane maps very well from directly above, which allows damaged areas to be outlined and measured.
Two papers in the journal Agronomy frame these uses well. A team led by Marcelo Rodrigues Barbosa Júnior published in 2021 a study on mapping gaps in sugarcane fields from drone RGB imagery, varying pixel size (3.5, 6.0 and 8.2 cm) and plant height (0.5 to 1.7 m): the best performance, with an error of about a centimetre and a half, comes at a 3.5 cm pixel over still-short cane — hence the title, the lower and earlier the flight, the more accurate (see the study on Google Scholar). The following year, a team led by Romário Porto de Oliveira published a method for predicting sugarcane biometric parameters — tiller number, plant height, stalk diameter — from drone multispectral images acquired every 40 days over five varieties, using multiple linear regression and random forest, with the blue, green and near-infrared bands proving the most predictive (see the study on Google Scholar). Those are precisely the variables that feed a pre-harvest biomass estimate.
The cane cycle sets the flight calendar — and the tool's limits
Cane is a multi-year crop: you plant setts, you harvest, the stool regrows, and you harvest again. Replanting is usually only considered after five to eight harvest cycles, depending on ratoon vigour, since yield declines cut after cut. That trade-off — one more ratoon, or replant? — is exactly what a drone illuminates best, because it is the only affordable way to quantify the gap rate across a grower's whole set of plots rather than along a 100-metre sampling line. The useful calendar therefore settles on three windows: after the cut, with short cane, to count gaps and judge regrowth; at full growth, for the vigour map and a targeted nitrogen top-up; before harvest, for the biomass estimate that helps schedule deliveries to the weighbridge during the milling campaign — July to December in Réunion, where the 2025 season opened on 22 July at Bois-Rouge for the north and east, and 28 July at Le Gol for the south and west.
Be clear about what a drone does not do. It does not measure sugar content: that is determined in a laboratory on a sample taken at delivery, the historic remit of CTICS in Réunion since 1952 and of the CTCS bodies in Guadeloupe and Martinique, and it is what sets the price paid to the grower. It does not replace the mill weighbridge: a biomass estimate is an order of magnitude for logistics, never a commercial figure. It does not diagnose disease on its own: a spectral anomaly flags an area to visit, not a named pathogen — identification stays with the technician, sample in hand. Finally, relief, the trade winds and airport control zones (Roland-Garros and Pierrefonds in Réunion, Pôle Caraïbes in Guadeloupe, Aimé-Césaire in Martinique) genuinely constrain flight windows: on this terrain, planning matters as much as hardware. On sensor choice, our comparison of a multispectral or RGB camera on an agricultural drone lays out the trade-offs, and the biomass and height estimation by drone logic developed for grassland applies, with ground calibration, to cane several metres tall.
Who commissions this kind of mission, method and 2026 prices
The buyers are first of all medium and large growers and grower groups, who pool a flight across several neighbouring farms to spread the travel cost. Then come the sugar mills and rum distilleries, interested in biomass estimates at supply-basin scale and in steady deliveries through the campaign; the cooperatives and agricultural development bodies, to target replanting programmes and fertilisation advice; and insurers and loss adjusters, for lodging assessments after a cyclone. On that last point the method is the one described in our guide to crop insurance and the claim file after a climate loss: a dated, georeferenced orthophoto, flattened areas measured rather than estimated, and comparison with a prior state where a baseline flight exists. This kind of mission is distinct from drone spraying in banana plantations against black Sigatoka, which falls under a far heavier spraying framework: here we observe and measure, we apply nothing.
Ranges observed in 2026, excl. VAT, for work carried out by a locally based remote pilot:
| Service | Range (excl. VAT, 2026) |
| Multispectral vigour-mapping flight, grouped plots up to 50 ha | €500 to €1,200 |
| Ratoon assessment and gap counting after cutting, per 20 ha block | €350 to €800 |
| Post-cyclone lodging assessment, dated orthophoto and measured flattened areas | €600 to €1,500 per farm |
| Campaign monitoring, 3 to 4 passes with biomass estimation | quote-based, by area and plot fragmentation |
These amounts cover the flight, photogrammetric processing, index computation and delivery of georeferenced maps. They exclude sugar-content analyses, soil sampling and the ground calibration measurements needed to turn an index into tonnage — agronomic work to be done with the sector's technical staff. Plot fragmentation weighs on price more than total area does: ten scattered two-hectare blocks cost more than twenty hectares in one piece. For a grower, a grower group, a sugar mill or a distillery, request a quote stating the island, the area, the number of blocks, their altitude and the goal (vigour, gaps, biomass or damage assessment).