C-DRONE GUIDE · 13 AUGUST 2026
Multispectral or RGB camera on an agricultural drone: differences, uses, price
Two drones can fly over the same wheat field on the same day and report two radically different things: one a colour-accurate, high-resolution aerial photo ready for measuring and counting; the other a set of spectral bands invisible to the eye that reveals a nitrogen deficiency one to two weeks before it shows on the ground. Between a standard RGB camera and a calibrated multispectral sensor, the choice is not about drone range or available budget — it is about what the mission needs to detect. Here is what each sensor actually sees, the uses where RGB alone is plenty, those where multispectral becomes necessary, and the prices charged in 2026.
Published on 13 August 2026, reviewed on 13 August 2026 — regulations in force as of August 2026.
RGB and multispectral: two sensors, two logics
An RGB camera mounted on a drone works like any regular camera: it captures visible light in three broad bands — red, green, blue — and produces an image directly readable by the human eye, at high resolution (often 20 to 45 megapixels on professional sensors). It is the default sensor for almost all drone missions: photogrammetry, counting, visual inspection, video.
A multispectral sensor works differently: alongside those visible bands, it adds narrow bands beyond red — the red-edge and near-infrared (NIR) — invisible to the naked eye. Spatial resolution is lower and the unit costs significantly more, but each band is radiometrically calibrated, often using a reference panel photographed before the flight or an onboard irradiance sensor. That calibration is what makes it possible to compare two flights taken on different dates or under different light conditions — an essential prerequisite for tracking a crop's evolution over time, something a plain RGB photo cannot reliably do.
What an RGB camera alone already covers
A large share of agricultural missions need nothing more than an RGB camera. Photogrammetry — orthophoto, digital surface model, plot area measurement — relies on high visible-light resolution, not infrared. Counting crop emergence and seedling density is done by visually detecting plants, row by row. High-contrast visual damage — hail, game, lodging — reads directly on a colour image.
RGB also allows simplified vegetation indices to be calculated without any infrared band. A landmark study by Gitelson, Kaufman, Stark and Rundquist published in 2002 (Remote Sensing of Environment) showed that an index built from visible bands alone — VARI, the Visible Atmospherically Resistant Index — estimates a field's vegetation fraction with an error below 10%, across a wide range of atmospheric conditions, without requiring any radiometric calibration (see the study on Google Scholar). To assess a cover rate, weed encroachment or uneven emergence, a plain RGB drone is therefore largely sufficient — and costs two to three times less to mobilise than a multispectral flight.
When multispectral becomes necessary
Near-infrared and red-edge do not measure a leaf's colour but its internal structure — mesophyll density, chlorophyll content — physiological parameters that often degrade several days before any yellowing or wilting becomes visible to the eye. That head start is what justifies the extra cost of multispectral as soon as the mission aims to detect stress before it shows visually: water stress in viticulture, nitrogen deficiency for a prescription map on wheat or rapeseed, or tracking grape ripeness at harvest.
A study by Candiago, Remondino, De Giglio, Dubbini and Gattelli published in 2015 (Remote Sensing, MDPI) compared several indices calculated from drone-mounted multispectral images — NDVI, GNDVI, SAVI — over vineyard and tomato plots, to finely map crop vigour at the scale of a single vine or row (see the study on Google Scholar). That level of detail — telling apart contrasting vigour zones within a plot that looks uniform at a glance — is precisely what an RGB sensor, however high its resolution, cannot deliver: the green of a slightly deficient leaf stays, to both the eye and a visible-light sensor, very close to that of a healthy one.
Choosing the right sensor for the job, 2026 prices
Three questions are usually enough to decide: does the mission aim to measure a geometry (area, count, orthophoto) — RGB; to detect stress before it is visible (nitrogen, water, early disease) — multispectral; or to compare several dates across the season to establish a trend — multispectral, for the radiometric calibration that makes flights comparable with each other. Many farms combine both over a year: one or two RGB flights for counting and a general survey, one or two multispectral flights timed to key growth stages (stem elongation, flowering, véraison) for input variable-rate management.
Ranges observed in France in 2026 (excl. VAT):
- RGB flight (orthophoto, count, visual vigour mapping, up to 20-30 ha): €400 to €900, depending on the resolution requested.
- Calibrated multispectral flight (NDVI/NDRE map, prescription map): €8 to €20 per hectare, with a minimum call-out package of €300 to €500, tapering beyond a hundred grouped hectares.
- Season subscription combining both sensors at several key stages: quoted case by case.
The extra cost of multispectral is worth it as soon as the data drives a decision — adjusting a dose, flagging a plot to monitor, deciding a selective harvest — not for a simple visual survey. This service draws on our agricultural drone offer: request a quote stating the crop, the area and the mission's objective to receive a proposal with the right sensor.
Frequently asked questions
Can a consumer drone with an RGB camera be enough for a vigour map?
For a visual estimate of vegetation cover or plot uniformity, yes — an index like VARI is enough. To detect stress before it is visible, or to compare several flights over time, no: that requires the infrared bands and radiometric calibration of a dedicated multispectral sensor.
Does multispectral replace the thermal sensor for water stress?
No, the two are complementary: thermal measures canopy temperature, linked to stomatal closure once water stress is advanced; multispectral (particularly red-edge) detects chlorophyll and leaf-structure changes, often earlier. Many vineyard missions combine both sensors on the same flight.
Can multispectral images be processed in-house, or is a service provider needed?
The flight and radiometric calibration require a dedicated sensor and a pilot trained in the protocol (reference panel, stable light conditions). Processing — stitching, index calculation, export to a spreading console — then requires dedicated photogrammetry software; most farms hand the whole job to an equipped provider rather than invest in equipment and training for a handful of flights a year.