C-DRONE GUIDE · 14 AUGUST 2026
Drone survey of a road accident scene for automotive experts: method, legal framework, price
A braking mark fades within hours under rain or the tyres of passing traffic; a shard of glass or a piece of broken bumper disappears the moment the road reopens to traffic. Yet it is precisely in these fleeting traces that the speed, trajectory and point of impact of a road accident can be read — the information an automotive expert, mandated by an insurer, a lawyer or a court, must reconstruct, sometimes several days after the event. A traditional survey with a measuring wheel or a total station takes hours and closes a lane; a drone, by contrast, captures a set of georeferenced photos in a few minutes, from which photogrammetry software produces a 3D model usable for reconstruction. Here is what this service can do, its limits next to the role of law enforcement, the regulatory framework that applies, and the prices charged in 2026.
Published on 14 August 2026, reviewed on 14 August 2026 — regulations in force as of August 2026.
Why the real problem is the time window
At an accident scene, law enforcement prioritises two goals that often pull against each other: securing the site and reopening the road as fast as possible. A full metric survey of the ground marks — braking, skidding, road-surface scoring, the final resting position of the vehicles — takes time with conventional tools, while every minute the road stays closed adds to congestion and the risk of a secondary crash. The result: on most material-damage or minor-injury accidents, the road reopens within a few hours and the marks vanish with it, before an expert outside the initial response can even get there.
Two situations change that. The first, and the most common route for a private drone pilot: a serious or fatal accident places the site under judicial investigation, which sometimes extends the closure or the preservation of evidence beyond the first response — a window in which a court-appointed expert can bring in a drone operator, working in coordination with the officer in charge of the scene. The second, more common in practice: the expert is not surveying fresh marks at all but the geometry of the site itself — the bend's profile, the road's camber, sight distance, the exact position of signage — data that does not fade with the rain and can be captured weeks after the accident to feed a trajectory-reconstruction program.
Photogrammetry, a faster alternative to the total station
The traditional survey method — measuring wheel, total station or GNSS receiver — remains accurate, but sequential: every point measured on the ground requires walking to it, sighting it, logging it. On a junction or a multi-lane stretch of motorway, the job can take several hours. A drone flips that logic: a ten-to-twenty-minute photogrammetric flight, anchored to a handful of GNSS-georeferenced ground control points, captures the entire area in a single pass. Processing then produces an orthophoto and a dense point cloud from which every needed measurement — distances, angles, the relative position of debris — is extracted back at the office, at a centimetre-level accuracy comparable to a total station, with no need to return to the scene for a point that was missed.
A study by Liu, Li, Zhang and Mao, presented in 2019 at the International Conference on Sensors, Signal and Image Processing, sets out exactly such a protocol for reconstructing an accident scene through drone photogrammetry: aerial imagery of the site, then extraction of a usable 2D and 3D model to place vehicles and ground marks with the precision an expert assessment requires (see the study on Google Scholar). A broader review of the scientific literature on the topic, published in 2023 in IEEE Access by Gohari, Ahmad, Abdul Rahim, Elamin, Gismalla, Oluwatosin, Hasan, Ab Latip and Lawal, surveys drone use in road accident management from 26 studies shortlisted out of 147 identified: it confirms that three-dimensional scene reconstruction is by far the most documented application in the academic research (see the study on Google Scholar).
Coordination with law enforcement and the regulatory framework
Since the January 2022 law on criminal liability and domestic security, the police and gendarmerie can deploy their own drones to secure an operation, subject to a préfecture authorisation limited in duration and purpose. A private operator commissioned by an expert, an insurer or a lawyer serves a different role: it never replaces the unit in charge of the scene, it works alongside it — most often once the site has been secured, or with the explicit authorisation of the officer in charge if the flight takes place during the response itself. No take-off should happen without prior coordination, if only to avoid any risk of collision with a state aircraft already on site.
On the administrative side, a survey above a busy public road, with residents, vehicles and sometimes pedestrians nearby, generally falls under open-category subclass A2 or A3 depending on the distance actually kept from third parties, or even the STS scenario of the specific category for an operation closer to people present; in a built-up area, the standard prior notification to the préfet applies under the usual rules — which is exactly why the deferred, plannable geometry survey, free of any time pressure, remains the easiest option for most expert practices to organise. The images captured, which can show number plates or identifiable people, are subject to the same GDPR framework as any other drone mission: use limited to the purpose of the expert mandate, with blurring before anything leaves the case file.
Deliverables and 2026 prices
Three deliverables cover most of what an automotive expert or an insurer needs: the georeferenced orthophoto of the site, with scale and distances directly measurable; the 3D point cloud, importable into trajectory-reconstruction software to place vehicles and obstacles back in their real environment; and a set of time-stamped photos with documentary value, comparable in principle to the images a commissaire de justice would have recorded for another type of dispute. As with construction expert-witness work, these deliverables are designed to be appended to an expert report, not to circulate as-is.
Ranges observed in France in 2026 (excl. VAT):
- Deferred geometric survey of the site (bend, junction, stretch of road, no fresh marks to capture): €300 to €700 depending on area and the number of ground control points laid.
- Full 3D model for reconstruction software (dense point cloud, centimetre-level accuracy): €700 to €1,500 depending on scene complexity.
- Urgent intervention on fresh marks, coordinated with law enforcement: a call-out surcharge of €200 to €400 on top of the survey fee.
This service draws on our surveying and photogrammetry offer, and can be cross-checked where relevant with a road signage inventory when the dispute concerns a design defect. Request a quote stating the nature of the accident, the mandate (insurer, lawyer, court-appointed expert) and the time available before the site reopens to traffic.
Frequently asked questions
Does the drone replace the survey done on site by the police or gendarmerie?
No: the unit in charge of the scene remains solely responsible for securing it and, where relevant, for its own aerial survey. A private operator commissioned by an expert works alongside it, most often once the site is secured or with the explicit authorisation of the officer in charge, never in competition with law enforcement on scene.
What accuracy can be expected for a trajectory reconstruction?
With GNSS-georeferenced ground control points, drone photogrammetry reaches centimetre-level accuracy, comparable to a total station — largely sufficient to place a point of impact, a braking mark or a vehicle's final resting position in reconstruction software.
Who can commission this type of survey?
An insurer, a lawyer, an automotive expert appointed by mutual agreement, or a court-appointed expert. The mandate must be formalised, not least because the images captured may show identifiable people or number plates, subject to GDPR and to use strictly limited to the purpose of the assessment.