C-DRONE GUIDE · 20 JULY 2026
Vehicle fleet counting by drone for car lots: method, accuracy and price
Multi-brand dealership, reconditioning platform, rental agency or port depot: counting a lot of several hundred vehicles on foot ties up two staff for half a day, an exercise repeated for every insurance or financing audit. A drone photographs the whole lot within minutes and turns it into an automated count, handled as data rather than a paper tally. Here is how the method works, its real accuracy, the applicable regulatory framework, and what it costs in 2026.
Published on 20 July 2026, reviewed on 28 July 2026 — regulations in force as of July 2026.
Counting hundreds of vehicles on foot: an exercise that no longer scales
Counting the vehicles on a storage lot remains, in most businesses, a manual job: one or two staff walk the rows on foot or by golf cart, record plates or serial numbers on a tablet, then cross-check against the computer inventory. On a lot of three hundred vehicles, the exercise ties up two people for half a day, and gets repeated every month, every quarter, or for every audit: fleet contracts and leasing companies require a dated, reliable stock count to justify the insured value.
Human error is not the only hidden cost: the time staff are tied up, the risk of double-counting or missing a vehicle on large, poorly lit lots with irregular lanes, and above all the impossibility of freezing a full inventory at one instant — vehicles keep moving while the team counts. A drone flyover answers exactly that last point: within a few minutes the whole lot is photographed at the same moment, giving a faithful snapshot of the real stock, with no drift between the start and the end of the count.
How the drone counts: grid flight, orthophoto and automatic detection
The mission runs as an automated grid flight (waypoints), at a constant altitude of 40 to 80 m depending on how dense the lot is and the resolution needed, with heavy overlap between shots. The images are stitched by photogrammetry into a single rectified, georeferenced orthophoto of the lot: every vehicle appears to scale, at its real position, comparable from one flight to the next. It is the same technical building block detailed in our guide to drone photogrammetry and 3D modelling, applied here to counting rather than volume surveying.
On this orthophoto, an object-detection algorithm automatically spots each vehicle by its shape and cast shadow, producing a total count and, if the lot is zoned (by brand, by status — awaiting prep, in transit, for sale), a count per zone by cross-referencing the polygons drawn on the plan. The delivered report usually pairs the annotated orthophoto (each detected vehicle boxed and numbered) with a summary table per zone, directly usable to reconcile the fleet-management system's inventory.
Real-world accuracy, blind spots and verification
Automatic detection reaches a high recognition rate on an open lot, with vehicles spaced apart and not overlapping — the most common case on a dealership or rental lot. It degrades in three specific configurations: vehicles parked very tightly in several rows with no visible gap between them, vehicles under a canopy, under a cover or underground — invisible from the sky by nature — and areas of dense cast shadow late in the day that blur the outlines. A serious operator therefore states a realistic detection rate, in the range of 95 to 99% on an open lot, rather than a round 100%, and systematically offers a cross-check on a representative sample before the report is validated.
For covered areas (hangars, showrooms, underground car parks), the drone does not replace a human count or a ground-level camera pass: it documents what is visible from the sky, and the report states that explicitly rather than implying full site coverage. That transparency about the method's limits, more than the algorithm's raw performance, is what sets a reliable operator apart on this kind of mission.
Regulatory framework: industrial, port and airport-adjacent lots
Vehicle lots are frequently located right next to an airport, a port or a classified industrial zone — an environment where the Géoportail drone map often flags a regulated zone or a reduced height ceiling unrelated to the site's own activity. That is the first thing to check before booking any mission, as we cover in our guide to the Géoportail drone zone map. If the lot is in a built-up area, professional flight there has been allowed in daytime open-category flight since 1 January 2026, with no overflight of people, after a prior préfecture notification with ten working days' notice — see our guide to flying a drone in French cities.
The pilot must in every case remain a registered operator on AlphaTango and be covered by aerial third-party liability insurance meeting Regulation (EC) No 785/2004, two points to check systematically before the mission: see our guides to AlphaTango registration and professional liability insurance. An open lot, with no staff or public in the immediate vicinity during the flight, falls under the standard open category in the vast majority of cases: our guide to open versus specific category covers the rare cases that would require a heavier framework.
The price of drone vehicle counting in 2026
The price depends on the lot's surface area, the number of vehicles to detect and above all whether the mission is a one-off or recurring: a contract of regular audits (monthly or quarterly) costs noticeably less per visit than a single intervention, since the pilot already knows the site and how the zones are laid out. Ranges observed in France in 2026:
| Mission | Observed price (excl. VAT) |
|---|---|
| One-off count of a lot under 300 vehicles (orthophoto + report) | €250 to €450 |
| One-off count of a large lot or port depot (300 to 1,500 vehicles) | €450 to €900 |
| Recurring contract (monthly or quarterly), per visit | €150 to €350 |
| Zoning by brand or status as an add-on (detailed report per polygon) | €80 to €200 |
For comparison, this service sits close, in its pricing logic, to drone stockpile volume measurement in quarries or civil works: the same principle of repeated aerial photography turned into usable data, applied here to counting units rather than calculating volumes. Many multi-site rental fleets and dealership groups now opt for an annual contract covering all their depots, which smooths the budget and simplifies coordinating flights from one site to the next.
Frequently asked questions about drone vehicle counting
Is automatic counting 100% reliable? No: expect a detection rate in the range of 95 to 99% on an open lot, with a cross-check on a sample before the report is validated. Covered, tarped or underground vehicles are not visible from the sky.
Does the lot's activity need to stop during the flight? No, the flyover takes a few minutes with no interruption to activity: vehicle movements, deliveries and departures carry on as normal, the photo simply reflecting the lot's state at the moment of the pass.
Can two counts be compared over time? Yes: that is even the main point of a recurring contract — cross-referencing two georeferenced orthophotos lets you check arrivals and departures between two passes, alongside the computer inventory.
Can the drone read number plates? Technically yes at low altitude on a sample, but that is not the mission's usual purpose: counting by zone (brand, status, location) covers the vast majority of audits, and flying lower sharply slows down coverage of the lot.