C‑DRONE
Aerial view of hillside vineyards under a dramatic sky

C-DRONE GUIDE · 30 AUGUST 2026

Unexploded Ordnance Survey Before Earthworks: What a Drone Adds (Magnetometry, LiDAR) and What It Does Not Do

A developer signs for thirty hectares north of Compiègne, or a local authority launches the development zone meant to host its future logistics hub in the Somme. Then somebody opens a 1918 map, and the plot turns out to sit exactly on a former front line. From that point, the earthworks schedule no longer depends only on excavators : it depends on an unexploded ordnance survey, on how long it takes to obtain, and on the very real risk of finding a shell under a bucket. A drone has no authority to remove anything from the ground — that is the job of France's civil security bomb disposal service and of licensed ordnance clearance contractors. But it is genuinely useful upstream, in two ways : reading the micro-relief inherited from the fighting, and carrying a magnetometer over ground nobody wants to walk on until they know what it holds. Here is exactly what it brings, where its limits lie, and the prices.

Published on 30 August 2026, reviewed on 30 August 2026 — regulations in force as of August 2026.

Why an unexploded ordnance survey comes before excavation

Across much of northern and eastern France, the subsoil still carries the physical trace of two wars. After 1918, the French war damages act of 17 April 1919 let the State buy up land judged impossible to return to cultivation : this was the zone rouge, covering 178,511 hectares as of 6 April 1919, from the Nord and Pas-de-Calais coast to the Vosges. Progressive declassification cut it to 48,820 hectares by 1 January 1927 — which means the large majority of that land was returned to farming and to development, cleared to 1920s standards. Many of today's development projects sit precisely there, on plots declassified a century ago. Add to that the sites bombed between 1939 and 1945 : rail junctions, ports, marshalling yards, factories, airfields, whose redeveloped footprints host a great many contemporary urban projects.

The pace of discoveries is anything but theoretical. France's civil security bomb disposal service collects and destroys several hundred tonnes of explosive devices every year ; a French Senate information report published in 2024 described a network of 26 field operational units, including 22 bomb disposal centres, staffed by 324 deminers, and already under strain because its storage depots are filling up. In other words : the question is not whether French soil still holds ordnance, but where, at what density, and at what cost to a construction schedule.

On the occupational safety side, the framework exists. Decree no. 2005-1325 of 26 October 2005 sets the safety rules applying to ordnance clearance works opened within a building or civil engineering project. It requires a pyrotechnic safety study, drawn up by the contractor holding the works, in two parts : the first covering preparation and the diagnostic phase — assessing the degree of contamination by detection over limited areas —, the second built around the families of items actually detected. That study must be approved by the labour administration before any works start. The file always opens with prior historical research : archives, trench maps, wartime aerial photography, bombing records. It is at that stage, and only that stage, that the drone comes in.

Reading micro-relief: photogrammetry, LiDAR and the digital terrain model

A century of ploughing, backfilling and reforestation does not erase everything. On many plots, shell craters, filled-in trenches, battery emplacements, collapsed shelters and 1944 bomb craters remain legible in the topography as depressions and rims a few tens of centimetres deep — invisible to the eye at ground level, perfectly visible on a shaded digital terrain model. That is exactly what the study by Hanne Van den Berghe, Wouter Gheyle, Birger Stichelbaut and co-authors, published in 2019 in Geografisk Tidsskrift-Danish Journal of Geography, documents : by combining a present-day LiDAR survey, a shell hole density map reconstructed for 1918 and a century of land-use change across the Ypres Salient, the authors show where crater landscapes survived and why — arable conversion erases them, permanent grassland and woodland preserve them (see the study on Google Scholar). That reading transposes directly to a French project footprint : it shows where the ground was most disturbed, and therefore where to concentrate the survey effort.

In practice the drone produces two complementary deliverables. A high-resolution orthophoto, on which soil colour differences and vegetation vigour often betray an old backfill or a filled-in trench, especially over low vegetation in raking light. And a digital terrain model isolating bare ground, an absolute condition for seeing thirty centimetres of micro-relief. Our guide to drone photogrammetry deliverables explains the difference between DTM and DSM, decisive here : it is the natural ground that speaks, not the canopy.

On open ground, photogrammetry is enough. Under forest cover — and a substantial share of the old battlefields was reforested in the 1920s — it sees only treetops : only a LiDAR survey penetrates the foliage and restores the ground surface. Our comparison guide LiDAR or photogrammetry sets out the choice criterion, and our guide to archaeological survey by drone describes the micro-relief analysis method, identical give or take a few centuries. To be clear : this work identifies landforms, not devices. A mapped crater says nothing about what it holds ; it says only that a shell detonated there, which is rather good news, and that the sector took intense fire, which is much less so.

Drone-borne magnetometry: what it actually measures

The second contribution is of a different nature. A magnetometer slung beneath the drone on a tether measures the intensity of the Earth's magnetic field along the flight lines ; a buried ferrous mass distorts it locally, and that distortion — the anomaly — shows up on the resulting map. The appeal in an ordnance context is immediate : you map without anyone setting foot on the plot, and you cover in a few hours an area that would take a ground detection crew days.

The orders of magnitude are documented. A study by Mick Emil Kolster, Mark David Wigh, Arne Døssing and co-authors, published in 2022 in Remote Sensing, surveyed a 600 m by 100 m footprint — six hectares — in roughly fifty minutes of cumulative flight at about ten metres per second, with the sensor towed some ten metres below the aircraft to keep it away from the drone's own electromagnetic interference ; the tested configurations recovered most targets, with noise floors on the order of tens of picotesla, but one of them, the horizontal gradiometer, performed markedly worse than the others (see the study on Google Scholar). First lesson : the sensor configuration and the way it is carried matter as much as the drone.

A study by Ignacio Ugarte-Goicuría, Diego Guerrero-Sevilla, Pedro and Javier Carrasco-García and Diego González-Aguilera, published in 2026 in the journal Drones, supplies the second and more important lesson for a client. At the San Gregorio manoeuvre area near Zaragoza, where eleven ordnance items had been buried at known coordinates, the survey was flown at 7 m and then at 2 m above ground level on a one-metre line spacing ; at low height, detection reached 100 % for ferromagnetic items of 60 to 155 mm calibre buried at less than 60 cm, with anomalies of only 2 to 18 nT, and this despite heavy ferromagnetic background noise typical of an active military range (see the study on Google Scholar). Note the trade-off : those figures depend on a two-metre flight height and a burial depth under sixty centimetres.

Hence a list of limits that belongs in the quote, not in the report :

Used for what it is — a prioritisation tool that steers resources towards the most anomalous sectors — drone magnetometry saves time and cuts exposure. Presented as a verdict on the presence or absence of ordnance, it becomes dangerous.

The boundary: what a drone does not and will never do

This has to be written flatly, because it is the source of the most expensive misunderstandings on the subject : a drone clears nothing. It does not dig, lift, neutralise or transport. Three distinct trades share this ground, and none of them is a drone pilot's trade.

The discovery of wartime ordnance is the bomb disposal service's business. Articles R. 733-1 and following of the French internal security code split responsibility between the civil security minister's services, operating on civilian land, and the armed forces ministry's services, competent on estates under their responsibility, in territorial waters and along the shoreline. Conduct on a site is invariable : do not touch, do not move, cordon off, evacuate, alert the police who task the deminers. No image, no magnetic map and no flight report changes anything in that chain.

The ordnance clearance of a construction site is the business of specialist contractors, under the decree of 26 October 2005 and the safety study approved by the labour administration. It involves grid-by-grid ground detection, borehole detection for depth, manual excavation on each anomaly, an effects zoning plan and evacuation procedures. The aerial survey substitutes for none of those steps : it prepares and documents them.

The ordnance risk assessment itself, finally, is analytical work : historical research, trench maps, wartime aerial photography, reconstructed fire density, the record of local discoveries, a conclusion on the degree of risk and on the investigations to be carried out. The drone's digital terrain model is a valuable input to that analysis — it shows the ground as it is today, at a resolution no archive provides — but the analysis belongs to the engineering consultant. For everything concerning site management once the ground is released, our guide to earthworks monitoring on a development zone and our guide to monitoring soil remediation works describe the drone's usual role : volumes, progress, traceability — the role of a witness, not an expert.

Method, flight constraints and 2026 prices

A properly framed mission runs in four stages. First a briefing with the engineering consultant and the client : exact footprint, vegetation state, metal fences and buried utilities, sectors already flagged by historical research, expected deliverable. Then the topographic survey — photogrammetry on open ground, LiDAR under cover — tied to GNSS control points so the model is georeferenced and can be overlaid on project drawings and wartime aerial photography. Then, if the programme calls for it, the magnetometric survey over the selected sectors : tight line spacing, the lowest flight height the terrain allows, an offset sensor, and a record of known noise sources so they can be discounted at interpretation. Finally the delivery : orthophoto, shaded DTM, georeferenced anomaly map, and a report that states explicitly what the measurement does not allow anyone to conclude.

Two constraints shape the schedule. In the air, a former front line is anything but neutral space : local aerodromes, military areas, sometimes permanent restricted zones. A check on Géoportail is mandatory before any quote, and our regional guides to drone work in Hauts-de-France and drone work in Grand Est set out the local specifics — including, around Verdun, the obvious precaution of staying airborne. As for season, micro-relief reads in low vegetation : late autumn, winter and very early spring give the best results in photogrammetry as in low-height magnetometry. A survey scheduled in July over tall scrub will deliver far less coverage for the same price. On very large linear footprints, the logic is the one described in our guide to the Seine-Nord Europe canal project : split it, prioritise, phase it.

2026 prices (excl. VAT), keeping the two families of service clearly apart :

A travel charge applies beyond a 30 to 50 km radius. And one clarification that should not need making : ordnance clearance itself is not a drone service and appears in none of these figures ; it is priced separately, by a licensed contractor, on the basis of the approved safety study. Our drone surveying and photogrammetry page presents the survey side ; request a quote stating the area, the vegetation cover, the intended period and who receives the deliverable (engineering consultant, client, clearance contractor).

Frequently asked questions

Can a drone survey replace the statutory unexploded ordnance survey?

No, at no point. The survey and the pyrotechnic safety study required by the French decree of 26 October 2005 are produced by specialist parties who carry the liability, and the study must be approved by the labour administration before works begin. The drone supplies two inputs to that file : a digital terrain model that objectifies the micro-relief inherited from the fighting, and possibly a magnetic anomaly map that helps prioritise sectors. It concludes nothing about the presence or absence of an explosive device.

How deep can a drone-borne magnetometer detect ordnance?

Far less deep than people assume, and flight height changes everything. On a Spanish experimental site where eleven ordnance items had been buried at known positions, a university team reported in 2026 a 100 % detection rate for ferromagnetic items of 60 to 155 mm calibre buried at less than 60 cm — but at a flight height of only 2 m, on a one-metre line spacing, and for anomalies of 2 to 18 nT. An item that is deeper, smaller, less ferrous, or flown over at seven metres instead of two may well produce no usable signal at all. That is why a drone magnetic map is read as a prioritisation tool, never as a clean-ground certificate.

What happens if ordnance is actually uncovered on site?

Nobody touches it, the sector is cordoned off and the police or gendarmerie are alerted immediately, who then task the competent bomb disposal service. The split of responsibilities is set by the French internal security code : the civil security minister's services for civilian land, the armed forces ministry's services for military estates and territorial waters. The drone pilot, like every other party on site, has strictly no role to play at that moment beyond stopping the flight and vacating the take-off area if it falls inside the safety perimeter.

Request a free quote

Also worth reading