A thermal drone pairs an infrared sensor with a stabilised aerial platform, turning invisible heat into a readable image from the sky — finding people at night, faults in electrical gear, moisture in roofs and animals in cover, none of which a normal camera can see.
- 8–14 µmThe band thermal works in
- 4 specsThat actually matter
- 0 lightRequired to see
- 0 wallsIt can see through
Key takeaways
- Every object above absolute zero emits infrared. Microbolometers read that emission and map temperature differences to pixels — no light needed.
- It sees people, animals, hotspots, moisture and heat loss through darkness and light smoke. It does not see through walls, glass or water.
- Four specs decide a payload: resolution, sensitivity, lens focal length and whether it is radiometric.
- Thermal detects; it does not identify. Faces and plates are the visible camera’s job.
On this page
How Thermal Imaging Works in One Minute
Every object above absolute zero emits infrared radiation, and how much it emits depends on its temperature. A thermal camera collects that radiation and maps the differences to pixels — brighter for warmer, or whatever the palette is set to.
Microbolometer sensors read this emission in the 8–14 µm band. Each pixel is a tiny membrane that warms slightly when infrared lands on it, and the readout measures the resulting change in electrical resistance. How uncooled VOx microbolometers work covers the detector.
No light is needed, because the target supplies the signal. Midnight and noon look the same to the sensor — which is the single property that makes thermal fundamentally different from every other camera on a drone.
Why that particular band? Because the atmosphere is transparent there and objects at everyday temperatures radiate most strongly there. Why thermal cameras live at 8–14 µm covers both facts.
What a Thermal Drone Can and Cannot See
The boundary is sharper than most people expect, and knowing it precisely prevents both disappointment and overclaiming.
| Can see | Cannot see |
|---|---|
| People, animals and vehicles in the open | Anything through walls, glass or water |
| Electrical hotspots and failing bearings | Faces, licence plates, text — any identifying detail |
| Moisture patterns under roof membranes | Colour, or anything about appearance |
| Building heat loss and thermal bridging | Objects at the same temperature as their surroundings |
| Solar panel faults | Subjects under solid cover or dense canopy |
| Targets in darkness and light smoke | Through heavy rain or dense fog at full range |
The two most consequential entries are on the right. No through-wall imaging: glass is opaque at these wavelengths and so is everything solid, which settles most privacy questions — see the privacy primer. No identifying detail: thermal renders a warm shape, and telling who that shape is requires a visible or low-light channel, which is why serious night payloads carry both — thermal vs night vision covers the pairing.
The Four Specs That Matter
Payload datasheets list dozens of numbers. Four of them decide almost everything.
| Spec | What it controls | Rule of thumb |
|---|---|---|
| Resolution | Pixels on target, therefore range | 256×192 entry, 640×512 professional |
| Sensitivity (NETD) | Smallest temperature difference visible | Lower is better; read the conditions with it |
| Lens focal length | Range versus area coverage | Narrow reaches further, wide searches faster |
| Radiometric or not | Whether you can quote temperatures | Cannot be added after purchase |
Deeper guides cover each: resolution, sensitivity, lenses and radiometry. The last one deserves emphasis for newcomers, because it is the only irreversible decision in the list — radiometry depends on factory calibration and no firmware update adds it later.
Two more worth knowing once the basics are settled: pointing accuracy, which caps usable zoom, and pixel pitch, which indicates core generation. What buyers should ignore first covers the specifications that consume attention without changing outcomes.
Missions That Justify the Investment
Thermal earns its place where the alternative is either dangerous, slow or impossible.
Finding people: search and rescue at night, where a warm body against cool ground is unmistakable and a visual search finds nothing.
Finding heat that should not be there: power line and solar inspection, where a failing connection or a faulty cell announces itself thermally long before it fails visibly.
Finding what is hidden: roof moisture and building envelope surveys, where wet insulation and missing batts have no visual signature at all.
Finding what is trying not to be found: security patrol and wildlife work, where camouflage and cover defeat the eye but not body heat. Each has a dedicated guide with payload recommendations.
Getting Started Sensibly
The most common beginner mistakes are buying the wrong resolution for the working distance, skipping radiometry and then needing it, and assuming the camera does the interpretation.
- Define the mission first: what has to be detected, at what distance, in what conditions.
- Choose resolution from working distance — under about 100 m, 256×192 is genuinely sufficient.
- Decide radiometric or not by whether your output will ever quote a temperature. This one is permanent.
- Check the payload fits the aircraft with margin — see payload weight versus flight time.
- Budget training alongside hardware; interpretation is where findings become defensible.
That last point is the one newcomers underweight. A thermal image is easy to produce and easy to misread — reflections look like hotspots, shiny metal under-reports its temperature, and solar loading mimics faults. Seven thermal imaging mistakes covers the recurring errors, and certification covers how to avoid them systematically.
Related reading
- How Uncooled VOx Microbolometers Work
- Radiometric Temperature Measurement
- Why Thermal Cameras Live at 8–14 µm
- 256×192 vs 640×512 Resolution
- Thermal Sensitivity (NETD) Explained
- Radiometric vs Non-Radiometric Thermal
- Seven Thermal Imaging Mistakes Operators Keep Making
- The Complete Guide to Drone Inspections
- MV-2M — 110 g dual-sensor micro pod
- MV-2P — 640×512 thermal at 130 g
- OP-90A — zoom, thermal and ranging
- Thermal Drone Payloads for Search and Rescue
FAQ
Do thermal drones work in daylight?
Yes — they read heat, not light, so daylight neither helps nor hinders detection directly. What daylight does change is contrast patterns: solar loading warms surfaces unevenly and can mask or mimic the anomaly you are looking for. This is why inspection timing matters so much more than time-of-day capability, and why roof and envelope surveys are evening work while irrigation surveys are midday work.
What does a professional thermal payload cost?
It varies several-fold depending on six drivers: thermal core resolution, radiometric calibration, zoom optics, stabilisation class, laser ranging and environmental sealing. Two payloads that look identical in a render can differ enormously because those items are invisible in a headline specification. Request the full configuration on every quote and compare configurations rather than resolutions.
Can a thermal drone see through walls?
No. Long-wave infrared does not penetrate solid material, and it does not penetrate ordinary glass either — glass is effectively opaque in the 8–14 µm band. A thermal camera reads the exterior surface temperature of whatever it can see directly. This is worth stating clearly and early, because most privacy objections to thermal drones assume a capability the physics simply does not provide.
Questions about the technology? Talk to our engineers — we reply within 2 business days.