C-DRONE GUIDE · 24 AUGUST 2026
Professional Drone Thermal Camera: Resolution, NETD and Radiometry — How to Choose in 2026
Two providers each offer a "drone thermal diagnostic" on the same building, and hand over two incomparable reports: one shows a simple colour-tinted image with blurred edges, the other a calibrated temperature map where every pixel reads in degrees and cross-references against an orthophoto. The difference almost never comes down to the pilot — it comes down to the sensor, and to three specifications few clients know to ask about before signing: radiometry, resolution and NETD. Here is what those words actually mean, why they weigh more on report quality than the price tag does, and how to check them before choosing a provider or equipping your own fleet.
Published on 24 August 2026, reviewed on 24 August 2026 — regulations in force as of August 2026.
Radiometric or not: the question that decides everything
A drone thermal camera turns long-wave infrared radiation (the LWIR band, 7 to 14 microns) emitted by any surface into an image, using an uncooled microbolometer sensor — the same technology, miniaturised, found in ground-based industrial thermal cameras. But not every thermal camera is equal from this first step onward: a non-radiometric camera, often bundled with consumer or entry-level drones, simply assigns a colour to each level of signal received — useful for spotting a visual contrast, unusable for a professional report. A radiometric camera, the only kind a serious provider uses for building or industrial thermography, calculates an actual temperature in degrees for every pixel, from the signal received, the target material's emissivity and compensation parameters (reflected apparent temperature, distance, humidity).
This distinction is not a marketing detail: only a radiometric image lets you plot a temperature profile along a façade, objectively compare two campaigns run months apart, or produce the figures required for a regulatory energy audit or an energy savings certificate file. Before any mission, the question to ask a provider is simple: is the file delivered a coloured visual JPEG, or a native radiometric file (often R-JPEG or TIFF format) usable in thermography software?
Resolution and NETD: the two numbers that determine what you can see
Two technical specifications, rarely highlighted on a sales sheet, determine what a thermal sensor can actually detect. Resolution — the number of pixels in the array, for instance 160×120, 320×256 or 640×512 — sets the size of the smallest defect identifiable at a given flight distance: at 30 metres of altitude, a 640×512 sensor resolves details two to three times finer than an entry-level 160×120 sensor, or covers the same roof in half as many passes for an equivalent ground resolution. On a large industrial building or a photovoltaic plant spanning several hectares, that gap directly changes the number of flights needed, and so the price of the mission.
| Array resolution | Typical use | Practical limit |
|---|---|---|
| 160×120 | Rough spotting, entry-level drone | Fine defects invisible beyond a few metres |
| 320×256 | Standard roof or façade inspection | Good price/detail trade-off for most missions |
| 640×512 | Professional thermography, large sites | Standard for specialised providers in 2026 |
The second number, NETD (Noise Equivalent Temperature Difference, expressed in millikelvins, mK), measures the smallest temperature difference the sensor can distinguish from electronic noise: the lower it is, the more the camera detects fine thermal variations — a 0.5 °C hot spot on a failing photovoltaic diode, a water infiltration that raises the surface by only a few tenths of a degree. Current professional sensors drop below 50 mK, against 100 to 150 mK on an entry-level model — a gap that, on a well-insulated roof where thermal anomalies stay subtle, is literally the difference between a defect caught and a defect missed.
A professional sensor is not enough: what a poorly prepared mission ruins
The best radiometric sensor on the market does not make up for a poorly prepared flight. A study by Linyuan Ma, Bing Xiong and Qingzhao Kong, published in 2024, tested under controlled conditions the effect of viewing angle, distance and direct sunlight on drone-based detection of tile debonding defects on building façades: the authors show that these three factors, far more than sensor resolution alone, sharply degrade diagnostic reliability when left unmanaged (see the study on Google Scholar). A flight carried out in full sun, with too oblique a viewing angle or a distance poorly matched to the sensor's focal length, can mask defects that the same aircraft would have revealed perfectly under good conditions.
An experienced provider builds these conditions into the flight plan as a matter of course: a sufficient indoor/outdoor temperature differential for building thermography, a dawn or evening slot to avoid solar reflection off metal and glazed surfaces, a viewing angle close to perpendicular to the inspected surface, and limited wind so as not to distort convective cooling. It is precisely this protocol, more than the sensor's spec sheet alone, that separates a usable thermography report from mere decorative imagery — worth checking upfront for any mission on a wind turbine, a roof ahead of a solar project, or an industrial façade.
Artificial intelligence changes how data is processed, not how it is captured
A thermography campaign across a large industrial site or a multi-site building portfolio quickly generates several thousand images to sort. A study by Qiwen Li and co-authors, published in 2024 in Infrared Physics & Technology, proposes a two-stage detection network, combining localisation and segmentation, to automatically identify debonding zones on drone-acquired thermal façade images (see the study on Google Scholar). This kind of approach, covered more broadly in our guide to artificial intelligence and automatic defect detection, does not replace the certified thermographer's eye that validates and interprets each anomaly — AI mainly serves to pre-sort, within a volume of data nobody could review image by image, the zones that warrant human review.
For a business client, this translates concretely into a shorter report turnaround on large volumes — a municipal roof portfolio, a solar plant spanning several hectares, a chain's network of buildings — without lowering the bar on capture quality: an algorithm trained on well-acquired radiometric images produces a reliable sort; the same algorithm applied to blurry or poorly exposed images only amplifies the error.
Choosing by mission: equipment, provider and 2026 prices
The right choice starts with the intended use. A company internalising an occasional visual-inspection capability — our guide compares both options — can make do with a drone fitted with an entry-to-mid-range radiometric thermal camera (160×120 to 320×256 resolution), around €3,000 to €7,000 for the complete aircraft in 2026 (excl. VAT). A specialised provider producing certified thermography reports for third parties — energy audits, insurance expertise, pre-works diagnostics — invests in a platform fitted with a 640×512 sensor with NETD below 50 mK, whose price climbs to €10,000-€20,000 (excl. VAT) depending on the brand and optics fitted.
On the service side, a standard thermography mission — a residential roof or a small commercial building — is most often billed at €400 to €900 (excl. VAT); an industrial site, a multi-building portfolio or a photovoltaic plant spanning several hectares rises to €1,500-€4,000 (excl. VAT) depending on the surface area and the level of report required, in the same range as the missions detailed in our sector guides. Before signing, three questions are enough to tell a serious provider apart: is the file delivered radiometric, what is the resolution and NETD of the sensor used, and was the flight scheduled within a weather and time window compatible with a reliable measurement? Request a quote stating the type of site, its surface area and the goal sought — a one-off diagnostic, a regulatory audit or periodic monitoring — to receive a proposal with the right sensor for the job.
Put it into practice
- Drone aerial thermal imaging: rates and cities covered from €500
- Aerial thermal imaging in Amiens Hauts-de-France
- Aerial thermal imaging in Orléans Centre-Val de Loire
- Aerial thermal imaging in Nancy Grand Est