C-DRONE GUIDE · 5 SEPTEMBER 2026

CAP satellite checks (area monitoring system): what a drone survey adds on an amber or red-flagged parcel, 2026 prices

CAP subsidy checks are, for the most part, no longer decided by an inspector turning up in your yard. Since the 2023 campaign, an automated system reviews every declared parcel from Sentinel satellite imagery and returns a green, amber or red flag in Telepac, France's CAP declaration portal. In 2026 the first flags appeared on 18 June, with updates in early July, early August and early September — and one deadline that shapes everything else: 20 September, the last day to amend a declaration without penalty. Faced with a red flag, the farmer's question is always the same: how do you prove the parcel really was in the declared state, when the algorithm saw something else? A dated aerial survey can help — but not in any form, and not at any moment. Here is what it genuinely documents, and, just as importantly, what it does not replace.

Published on 5 September 2026, reviewed on 11 September 2026 — regulations in force as of September 2026.

What the area monitoring system is, and why on-the-spot checks became rare

Checks on area-based CAP payments now rest on a mechanism written into EU law: the area monitoring system, one of the components of the integrated administration and control system listed in Article 66(1)(c) of Regulation (EU) 2021/2116. It is Article 70 of that same regulation that requires the paying agency to put the system in place, with paragraph 2 providing for an annual quality assessment. The detailed rules were set by Implementing Regulation (EU) 2022/1173 of 31 May 2022, which among other things provides that the Commission supplies the satellite data free of charge to the competent authorities or to the service providers they mandate.

In France the system is known as the real-time area monitoring system, or 3STR. Parcel by parcel, it verifies that the declared land cover is real and that agricultural activity is taking place, from automatically analysed satellite imagery. The raw material is the European Copernicus programme's Sentinel satellites: optical Sentinel-2 passes over a given point every five days in the nominal two-satellite configuration (ten days with a single satellite), while radar Sentinel-1 sees through cloud cover with a repeat cycle of six to twelve days depending on the constellation available. It is that regularity — not image sharpness — that makes the system work: by stacking passes across a season, the algorithm reads a trajectory (emergence, mowing, harvest) rather than an isolated snapshot.

The result reaches you as traffic-light flags displayed in Telepac. A green flag means the declared cover is deemed compliant; an amber flag means it is not yet possible to conclude on compliance — analysis continues; a red flag means the cover is not considered compliant. For the 2026 campaign, the first flags were displayed on 18 June, at that stage green and amber only, reflecting the situation at the time of calculation in mid-June; updates were scheduled for early July, early August and early September 2026. That sequence follows the cropping calendar: a winter crop qualifies earlier than a spring crop, and a summer crop later still.

Amber flag, red flag: what the administration actually expects

An amber flag normally calls for no action. It signals that the algorithm has not yet reached a conclusion — either the parcel needs further expert review, or it carries a late crop whose signature is not yet readable. Most resolve themselves at the following updates; the paying agency indeed put the share of parcels concerned below 3.8% in the trade press. The useful reflex is to check the flag has turned green at the next calculation, not to assemble a file in a rush.

A red flag, by contrast, demands a decision. Two routes open up, and they are not mutually exclusive:

• Amend your declaration in Telepac, through the "modify after submission" function, if the reported discrepancy is real — a miscoded cover, two parcels swapped, a badly drawn boundary. Under the right to make a mistake, that amendment carries no financial reduction and no penalty up to and including 20 September 2026, provided your declaration was signed and you have not been notified by the paying agency of an on-the-spot check covering the aid concerned. The scheme does not cover payment-entitlement transfers.
• Stand by your declaration and send the local agriculture department (DDT(M)) the context and supporting evidence that explains the discrepancy: seed invoices, delivery notes, operation schedules, contractor reports, dated photographs. The department forwards it to the paying agency for review, and the agency may decide on a field visit targeted at the criteria the automated system cannot verify.

The paying agency may also send you a geotagged photo request. Those pictures must be taken with a mobile phone through the free TéléPAC Géophotos app, which timestamps and geolocates them in an imposed format. It is a formal channel, distinct from any other document: if you receive such a request, it is handled inside the app, full stop.

Where the satellite hits its limits — and where a dated drone survey helps

A red flag is not a verdict: it is the output of an analysis chain whose limits are documented and published. A study by Blanka Vajsová, Dominique Fasbender, Csaba Wirnhardt, Slavko Lemajic and Wim Devos, published in 2020 in Remote Sensing and devoted precisely to the spatial limits of Sentinel-2 in the context of checks by monitoring, compared NDVI time series from Sentinel-2 (10 m pixels) with series from 3 m imagery across 867 fields under 0.5 hectare: for 10% of them, the Sentinel-2 data did not provide reliable evidence of the field's activity or state over the period studied. The authors derive limiting geometric criteria — at least eight full pixels inside the boundary, fewer than 60% of pixels lost after a half-pixel negative buffer (see the study on Google Scholar). In other words: on small fields, narrow fields, grass buffer strips or tree-lined blocks, the satellite is structurally at its resolution limit.

Detecting practices is harder still than detecting crops. A study by Mathilde De Vroey and co-authors, published in 2022 in Remote Sensing of Environment, built a mowing-detection method combining Sentinel-1 and Sentinel-2 time series for large-scale grassland monitoring, leaning on radar's regularity to fill the optical sensor's cloud gaps (see the study on Google Scholar). In the same vein, Viacheslav Komisarenko, Kaupo Voormansik, Radwa Elshawi and Sherif Sakr published in 2022 in Scientific Reports a deep-learning model detecting mowing events on Sentinel time series, fitted with a deliberate reject region: the model explicitly declines to decide when confidence is insufficient (see the study on Google Scholar). That reject region is, conceptually, what you see on screen as an amber flag.

It is precisely in that gap that a drone survey earns its place. At 2 to 5 cm per pixel, an aerial orthophoto shows the individual plant, the row, the drill line, the mowing trace, the boundary stake — objects a 10 m pixel cannot separate. The flight is triggered on demand, below the cloud base, on the day of the operation: where the satellite documents a five-to-twelve-day interval, the drone documents a date. And the resulting file carries its metadata: timestamp, coordinates, altitude, ground footprint. For a full comparison of the two technologies on agronomic ground, our guide to drones versus satellites in precision agriculture sets out the resolution and cost-per-hectare trade-offs.

What the drone does not do — and it needs saying plainly

A provider who sells you a drone flight as a guarantee that a red flag will disappear is selling you an illusion. Four limits need stating before any quote.

• The drone does not replace the geotagged photo. If the paying agency sends you a geotagged photo request, it is satisfied inside TéléPAC Géophotos, with a phone, in the stated format and within the stated deadline. No orthophoto, however sharp, substitutes for that channel. An aerial survey comes on top, as contextual evidence sent to the local agriculture department.
• The drone does not replace dialogue with the administration. Subsidy control is an administrative procedure: it turns on a declaration amended in time, a documented exchange with the case officer, a coherent explanation. The finest image file in the world does not make up for staying silent until 20 September.
• The drone does not travel back in time. A flight on 5 September documents 5 September. It does not prove a cover crop was established in April, nor that a meadow was mown in June. On a farm with recurring discrepancies, the only strategy that holds is to schedule flights against the cropping calendar, in advance, rather than react after the fact.
• The drone does not produce enforceable proof on its own. A timestamped orthophoto is contextual evidence, weighed as such by the case officer and, if it comes to it, by an administrative judge. Its value rests on traceability — intact flight metadata, georeferenced control points, a signed report describing the method — not on resolution alone. A serious provider hands you those elements; they do not promise you an outcome.

The practical rule: an aerial survey serves to objectify a field situation when the discrepancy concerns a visible physical reality (area genuinely farmed, presence of a landscape feature, state of a cover on the day of the flight). It is useless when the discrepancy concerns a declaration rule, a crop code or a past operation date. It is the same file-building logic described in our guide to what a drone survey adds to a crop-insurance claim file, transposed to a public counterpart.

Method, timing and 2026 price ranges

The ideal trigger is not the red flag: it is the cropping calendar. On a farm that has already seen discrepancies — late-mown meadows, catch crops, small landlocked parcels, buffer strips — the effective approach is to identify the five to ten at-risk parcels at the start of the campaign, then schedule one or two flights at the moments when evidence matters: just after a cover crop is established, in the days following a mowing, at the emergence of a spring crop. The minimum useful deliverable is a dated georeferenced orthophoto, together with a method note recording flight time, height, ground resolution and the raw metadata retained. A multispectral pass adds nothing to an administrative file about whether a cover is present: RGB is enough and costs less.

If the red flag is already there, the order of operations matters more than the flight itself: contact the local agriculture department to understand the exact reason for the alert, decide between amending and standing by the declaration, answer any geotagged photo requests within the deadline, and add the aerial survey only if it documents a reality still visible on the ground. On a dispute about eligible area or landscape-feature length, by contrast, the flight retains its full value at any point in the campaign: geometry does not fade — the subject of our guide to hedgerow mapping for CAP and conditionality requirements.

On budget, there is no published price list for this still-new type of mission. The ranges below are extrapolated from comparable agricultural services actually invoiced in France, in euros excluding VAT, and remain to be confirmed case by case:

• Survey of one or two parcels (up to around ten hectares), dated georeferenced RGB orthophoto: in the region of €250 to €500.
• Batch of 10 to 50 hectares, several blocks covered in a single trip: in the region of €500 to €1,100.
• Method note and file preparation for the administration (measured areas, dated plates, metadata): a further €200 to €500.
• Scheduled campaign across at-risk parcels, two or three passes spread over the season: sharing the travel cost lowers the unit price of each flight, all the more so when parcels are grouped.

One cost factor is routinely underestimated: responsiveness. A flight to be scheduled within 48 hours because a mowing is imminent is priced differently from one booked three weeks ahead. Anticipating costs less than reacting — true of a CAP file as of everything else. All our agricultural services are gathered on the precision agriculture by drone page; for a farm looking to frame a survey campaign across its sensitive parcels, request a quote stating the number of parcels concerned, their area, the nature of the reported discrepancies and the schedule of operations to be documented.

Frequently asked questions

Request a free quote

Put it into practice

• Drone security & surveillance: rates and cities covered from €600
• Security & surveillance in Tours Centre-Val de Loire
• Security & surveillance in Besançon Bourgogne-Franche-Comté
• Security & surveillance in Caen Normandie

Also worth reading

• Olive orchards and drones: water stress and yield forecasting, price
• Prune d'Agen and drones: water stress and ripeness, price
• Asparagus and drones: spotting weeds on the ridge before emergence, price