C-DRONE GUIDE · 4 AUGUST 2026
Noise mapping and source localisation by drone: acoustic camera, ICPE framework, price
Residents worn down by a wind turbine's whine, a quarry operator ordered to explain a night-time noise spike, a factory unable to say which of its ten machines breaches the regulatory threshold: pinpointing the exact source of a nuisance noise, on a site that produces several at once, is often harder than measuring it. A sound level meter gives a reliable value at one point; it says nothing about where the noise comes from. A drone-mounted microphone array, by contrast, flies over the site and produces a false-colour noise map overlaid on the video feed — enough to name the offending piece of equipment in minutes rather than days of ground investigation. Here is how this still little-known technique works, what it adds against industrial noise regulation, and what it does not replace.
Published on 4 August 2026, reviewed on 29 August 2026 — regulations in force as of August 2026.
The principle: a microphone array in false colour
A conventional acoustic camera pairs anywhere from a dozen to several hundred MEMS microphones arranged in an array with an optical camera: by measuring the tiny phase shift with which the same sound reaches each microphone, a beamforming algorithm computes the direction the noise comes from and renders it as a colour map overlaid on the image — much as a thermal camera does for heat, except it "sees" sound. Miniaturised and mounted on a drone, such an array can sweep a roofline, run along a row of machines or pass over a wind farm to map in minutes what a series of ground measurement points would take hours to cover.
The main technical challenge of this airborne approach is the drone's own noise — motors and propellers — which can mask or pollute the target source. A study by Yoon-Ju Go and Jeong-Guon Ih, published in 2021 in the journal Drones, proposes precisely a phased-microphone-array localisation method mounted on a drone, designed to separate the noise of the targeted source from the predictable noise generated by the aircraft itself (see the study on Google Scholar). It is this signal processing, more than the hardware itself, that makes the measurement usable in flight.
The legal frame: two logics, as with dust
Industrial noise in France follows a split close to that of dust monitoring: classified installations (ICPE) are subject to precise thresholds — the arrêté of 23 January 1997 caps the noise emergence, in regulated emergence zones (ZER), at 5 dB(A) during the day (7am–10pm) and 3 dB(A) at night (10pm–7am) — while non-classified sites and ordinary construction sites face no fixed regulatory threshold, but remain exposed to civil liability for abnormal neighbourhood nuisance and occasional municipal orders. The reference measurement, the one that holds up in an inspection, remains a standardised sound level meter following the method set out in standard NF S31-010 — a drone-mounted acoustic camera does not replace it.
It comes as a complement, with one advantage a fixed meter lacks: naming the precise source among several pieces of equipment running at once. That is a frequent need on wind farms, where blade aerodynamic noise draws recurring resident complaints alongside drone blade inspection campaigns, or on industrial sites tracked under an autonomous drone site surveillance plan.
Three typical missions
In practice, three situations come up most often:
- Responding to a neighbour complaint: a flight identifies, among several pieces of running equipment, the one actually dominating the reported noise — extraction fan, compressor, chiller unit — with a timestamped report usable in a reply to the town hall or the préfecture.
- Diagnosis ahead of soundproofing works: before fitting an enclosure or an acoustic screen, the mapping ranks the relative contribution of each source to target the investment rather than treating the whole site at the same level, a logic close to the one detailed in our guide on air quality monitoring by drone.
- Periodic tracking under an environmental monitoring plan: before/after passes once corrective measures are in place, to give the classified-installation inspectorate or a wind farm operator an objective read on their effectiveness.
In all three cases, the flight is kept far enough from the equipment to limit signal pollution from the drone's own noise.
What the technique does not replace
Three limits worth knowing before booking a mission. First, the reading is relative and cartographic, not a legally admissible dB(A) value: it shows where the noise dominates, not a certified measurement under standard NF S31-010. Second, the onboard array's range stays limited by how small the microphones are and by the safety distance imposed by the drone's own noise — flying too close to the source degrades the measurement rather than improving it. Third, wind disturbs an acoustic array far more than a thermal camera or a particle sensor: the usable weather window is narrower.
A drone-mounted acoustic camera therefore replaces no regulatory inspection procedure — it serves to steer the response ahead of a certified acoustician or engineering firm stepping in, exactly as a scouting flight guides a targeted human intervention rather than replacing it.
2026 prices and how to budget
Orders of magnitude observed in France in 2026 (excl. VAT):
- One-off scouting flight (responding to a complaint, a single pass, an annotated map): €400 to €800.
- In-depth mapping (several flight positions, source ranking, detailed report): €900 to €1,800 depending on site size and complexity.
- Periodic tracking (quarterly or annual, under an ICPE monitoring plan or after a recurring complaint): quoted case by case, typically tapering under a package.
This service draws on our security and surveillance drone offer: request a quote stating the site, the number of pieces of equipment potentially involved and the reason for monitoring (complaint, ICPE duty, diagnosis ahead of works).