PRF2Camera or remote sensor guided herbicide spraying

What it actually is

You apply your systemic herbicides on the entered land through camera or remote sensing technology that identifies the specific weeds needing control and automatically switches individual nozzles on to spray them directly, so the chemical goes on the weed rather than the whole field. Three kinds of application are allowed to stay as blanket passes: glyphosate to destroy a cover crop or create a stale seedbed, pre-emergence residual herbicides at or within days of establishing the crop, and pre-harvest desiccants on combinable crops.

The action page is unusually plain about what the technology can currently manage: docks and nettles in grass and leys with clover, all weeds in maize, crops on wide row spacings such as potatoes, and stubbles with green growth, with remote sensing using artificial intelligence to recognise the target weeds.

GOV.UK's stated aim is precise weed control that reduces the volume of herbicide applied while increasing biodiversity, improving water and soil quality, and supporting integrated pest management.

The action is rotational or static, so it can follow the sprayed ground around the rotation or stay put.

Source: GOV.UK, PRF2 action page, checked 23/08/2026.

What it pays, and what that comes to

The rate is in the panel at the top of this page. Multiplied out:

Source: arithmetic on the payment rate shown at the top of this page, which comes from Tilth's copy of the GOV.UK action catalogue and carries its own last-checked date. The 80 hectares is an illustration.

Is it worth it on your ground?

The payment is only half the money in this action, because unlike most actions on this site, the practice it pays for also cuts a bill you already have.

The chemical saving is real and measured. In the one independent on-farm trial we have found, across four maize fields in south west Germany in 2023, spot spraying guided by drone weed maps used 47 per cent less herbicide than the blanket pass when applied early, 44 per cent less when combined with a pre-emergence pass, and achieved weed control and yields equal to the blanket treatment. Put your own herbicide spend on the eligible ground against those percentages and the saving on chemical can stand comparison with the payment itself.

Source: Allmendinger and others (2024), Plants, 13(15), 2164, checked 23/08/2026. One trial, one crop, one country; treat it as an indication, not a promise, and note the timing caveat in the evidence section below.

The kit is the other side of the ledger. Camera-guided sprayers, retrofit spot-spray systems and drone mapping services all carry prices this page will not guess at, and the payment does not buy any of them. As with the variable rate action, the farms this suits are those that already have the technology, can hire it, or can bring in a contractor who runs it, and the contractor route is explicitly anticipated by the action's evidence list.

The eligibility test does some arithmetic for you. Only land receiving systemic herbicides for weeds the technology can recognise is eligible, so the entered area is exactly the ground where a saving is available, grassland with docks, maize ground, wide-row crops and stubbles, not the whole farm.

The kit reasoning is Tilth's own and adds no requirement beyond what the GOV.UK action page sets out.

Does it work, and what makes it work better

This section is what we have found in published sources about the practice this action pays for. It is not scheme guidance, it is not Defra's or the RPA's view, and it adds no requirement beyond what the GOV.UK action page sets out. We have named every source and given the date we checked it. What it means for your ground is your decision.

Conservation Evidence hold no assessment of camera guided or spot spraying, so this page rests on the agronomic literature, and the substantial piece we have found is recent and properly on-farm: Allmendinger and others ran a replicated trial across four commercial maize fields in south west Germany in 2023, comparing spot spraying, guided by high resolution drone maps of where the weeds actually were, against ordinary broadcast herbicide.

What they found. Early spot spraying, at the maize's 2 to 4 leaf stage, used 47 per cent less herbicide than the broadcast pass and achieved up to 86 per cent weed control efficacy, equal to the broadcast treatment. A programme of pre-emergence spray plus early spot spraying saved 44 per cent. Maize yields in the spot sprayed plots were equal to the broadcast plots, 44.6 to 50.7 tonnes per hectare of fresh weight across treatments with no significant differences.

And the caveat that matters most. Late spot spraying, at the 6 to 8 leaf stage, saved only 12 per cent of the herbicide and managed 20 per cent weed control against 76 per cent for the late broadcast pass, because by then the maize canopy shaded the later emerging weeds and the detection could not see them. The window in which this technology worked was the window in which the camera could see the ground.

Source: Allmendinger and others (2024), Agronomic and Technical Evaluation of Herbicide Spot Spraying in Maize Based on High-Resolution Aerial Weed Maps, An On-Farm Trial, Plants, 13(15), 2164, checked 23/08/2026.

What we have not found. An independent British trial of camera guided spraying in the grassland dock situation the action page lists first, which is where much of this action's uptake will sit; the results reported there so far are largely manufacturers' figures, which we do not cite. Nothing measuring PRF2 itself. And nothing on what the biodiversity, water and soil aims in the action's stated purpose actually gain from the reduced volumes, which are presumed benefits of less chemical rather than measured ones.

What the sources say makes it work better. One steer, and it is strong: go early. The entire difference between the trial's success and its failure was timing, 47 per cent saved with control intact early, against 12 per cent saved and control lost late, once the crop hid the weeds. Whatever the crop, the practice works while the target is visible.

Source: Allmendinger and others (2024), as above, checked 23/08/2026.

What follows is Tilth's own reading of how those findings sit against this action, and it adds no requirement beyond what the GOV.UK action page sets out.

This is a rare case where an action's design and the evidence point the same way. The action page's capability list, docks in grass, maize, wide rows, stubbles, is a list of situations where the weed stands visible against its background, and the one good trial says visibility is exactly what decides success. The practice is not so much spot spraying as spot seeing, and everything downstream follows from whether the camera had a clear look.

The honest gaps are British and financial. The measured result is maize in Germany; the biggest British use is docks in grassland, where the published independent evidence has not caught up with the machinery, and where your own before-and-after chemical invoices will be a better trial than anything in print. And the environmental half of the aim, water, soil, biodiversity, is inferred from less chemical rather than demonstrated, which is probably true and certainly unmeasured.

For a farm weighing it up, the sum is unusually concrete: your current systemic herbicide spend on eligible ground, cut by something like the measured percentages if your situations match the capability list, plus the payment, against the cost of the technology or the contractor who brings it. Every number in that sentence except the trial percentages is already on your farm.

What you actually have to do

Apply systemic herbicides on the entered land using camera or remote sensing technology that identifies the specific weeds requiring control and automatically switches individual nozzles to apply herbicide directly to them, in a way that can reasonably achieve the action's aim. Blanket applications are allowed only for glyphosate used to destroy a cover crop or create a stale seedbed, pre-emergence residual herbicides at or within days of crop establishment, and pre-harvest desiccants on combinable crops. Do this in each year of the three year duration.

Source: GOV.UK, PRF2 action page, checked 23/08/2026.

Where it can go, and what rules it out

Agricultural land below the moorland line that receives systemic herbicides for weed control which the technology can recognise: arable land with crops, temporary grassland under TG01, fallow arable land under FA01, improved and low input permanent grassland under PG01, and permanent crops with compatible codes. SSSIs need Natural England's consent first, and land with historic or archaeological features needs its HEFER. Rotational or static.

Source: GOV.UK, PRF2 action page, checked 23/08/2026.

Things worth watching

This section is Tilth's own reading of where this action is easy to get wrong. It is not scheme guidance, it is not sourced from Defra or the RPA, and none of it adds a requirement beyond what the GOV.UK action page sets out.

Enter the ground the camera can work, not the ground you spray. The eligibility wording, weeds the technology can recognise, is doing real work. A wheat field with a broad-leaved weed problem gets sprayed with systemics too, but if no current technology can pick those weeds out of that crop, it does not belong in the action. The page's own capability list is the honest guide to what qualifies today.

The exceptions list is exact, and everything else must go through the cameras. Glyphosate for covers and stale seedbeds, pre-em residuals at establishment, desiccants pre-harvest. A blanket dose of a selective on a growing crop is none of those, and the year someone reaches for one in a hurry is the year the as-applied maps stop matching the requirement.

Timing is compliance and agronomy at once. The trial's lesson, spray while the weeds are visible, is also the practical answer to staying inside the action: a spot system that cannot see the weeds does not control them, and a farm that then rescues the situation with a blanket pass has left the action on that ground. Plan the spot passes early.

Contractor kit is fine, contractor data is the point. If a contractor brings the camera sprayer, the as-applied maps showing which nozzles fired where come off their machine, and their invoice is named in the evidence list. Agree before the season that you get the maps, per field, in a form you can keep.

Watch the tank as well as the nozzles. The saving is in chemical actually applied. Spot systems still travel the whole field, so diesel and time move less than the chemical bill does, and anyone budgeting the saving as a percentage of the whole spray operation rather than the chemical line will overpromise themselves.

Keeping the evidence

Also Tilth's own reading. GOV.UK sets the requirements; what follows is a practical suggestion about record keeping.

GOV.UK's examples are the as-applied data, maps showing where individual nozzles activated, records of any permitted blanket applications with their dates, and contractor invoices where relevant.

The as-applied maps are the heart of it and they are generated by the kit, so the habit is simply keeping them: per field, per pass, filed where they can be found in year three. The blanket passes need the opposite discipline, a note at the time saying which exception each one was, cover destruction, pre-em at establishment, or desiccation, because a blanket application with no stated reason looks like a breach even when it was not.

Tilth records spray operations against the parcel with dates and products, lets you attach files and photographs to the job, and holds contractor invoices against the work, so the maps, the reasons and the paperwork stay together.

The part only your own ground can answer

What your systemic herbicides actually cost you on the eligible ground, because that number times the measured savings is this action's real value, and no page can know it. Your spray records hold it already: product, rate, area, price, season by season. Tilth keeps those records by field, which means the before is on file the day you start, and the after writes itself. Request an invite.