Thermal cameras defeat darkness, not weather. Long-wave infrared is absorbed and scattered by water in the air, so fog, rain and high humidity can cut a spec-sheet detection range by half or more — and dense fog can make thermal imaging almost as blind as a visible camera.
- 70–80%Of spec range in light haze
- 50%In steady rain
- UnreliableBeyond a few hundred metres in fog
- Clear, dryWhat DRI figures assume
Key takeaways
- Water droplets are close in size to LWIR wavelengths, so fog scatters thermal radiation efficiently.
- Rain adds a second penalty: it cools surfaces toward a uniform temperature, shrinking the differences you are trying to image.
- Published DRI figures assume a clear, dry atmosphere. Derate them deliberately rather than hoping.
- Close-range work keeps functioning — and roof moisture surveys after rain actually improve, because evaporative cooling deepens contrast.
On this page
Why Water Is the Enemy of LWIR
Water droplets in fog are close in size to LWIR wavelengths, which makes them efficient scatterers of thermal radiation. Radiation that should have travelled from target to sensor is redirected, and what arrives is a weaker, more diffuse signal.
Humid air also absorbs across the 8–14 µm band. The atmospheric window is a window rather than an open door, and water vapour is the main thing narrowing it. The result is lower apparent contrast: targets look dimmer and flatter the further away they are.
Rain adds a second, independent penalty that people often miss. Beyond attenuating the signal, it cools surfaces toward a uniform temperature — which shrinks the very temperature differences you are trying to image. A rained-on scene is both harder to see through and genuinely less thermally interesting.
Practical Range Planning
Published DRI figures assume a clear, dry atmosphere. Operators derate them, and the working numbers are reasonably consistent across the industry.
| Condition | Plan on | Practical note |
|---|---|---|
| Clear and dry | Spec range | The laboratory condition |
| Light haze | 70–80% of spec | Still comfortably usable |
| Steady rain | 50% of spec | Plus reduced scene contrast |
| Fog | Unreliable beyond a few hundred metres | Treat long-range planning as void |
| High humidity, no precipitation | Somewhere between haze and rain | Worse than it looks from the ground |
Compare range figures only when the sensor channel, target size, task definition and test conditions are stated. Johnson-based thermal DRI and visible-camera DORI are not interchangeable, and there is no universal conversion factor between them. Use the model-specific source document and validate the required task under your operating conditions.
Apply the derate to lane spacing, not just to expectations. A search pattern flown at clear-air spacing in fog covers less area than it appears to, and the gaps are invisible on the flight log.
Derate the plan, not the debrief. Adjusting expectations after a search found nothing is not planning. Tighten lanes and reduce standoff before launch, on the conditions you actually have.
What Still Works Well
Close-range missions keep working, and some improve.
Roof moisture surveys after rain actually improve. Evaporative cooling from the drying surface deepens the contrast between wet and dry sections — though the surface itself must be dry before the flight, or you are imaging surface water rather than subsurface moisture. The roof survey guide covers the timing.
Search work under drizzle remains viable at reduced standoff. The signal is attenuated but a person at 100 m is still a strong target; what is lost is the wide-area efficiency that made the aircraft worth launching. Tighten the pattern and accept slower coverage.
High-sensitivity sensors hold usable contrast longer as conditions degrade. A lower NETD means a smaller surviving temperature difference is still above the noise floor, which is exactly what atmospheric attenuation leaves you with — see the NETD explainer. This is one of the few situations where sensitivity buys operational capability rather than image quality.
And cold-weather work is often the inverse case: cold, dry air is excellent for thermal, and the constraint moves to batteries and crew.
The Payload Side
Weather affects the hardware as well as the imaging, and the two questions are separate. A sealed payload will survive conditions that make its imagery useless.
Check the ingress rating against the exposure, remembering that IP ratings cover static immersion and jets rather than rain at flight speed — what IP67 really promises covers the gap. The published environment specification across the line is −20 °C to +50 °C operating, ≤85% RH non-condensing, and that last qualifier is what a climb through a humid layer can breach.
Water on the lens is its own problem, distinct from water in the air. A film of water on a germanium optic becomes an emitting surface between the sensor and the scene, which shifts thermal accuracy as well as blurring the image. There is no software correction for it.
And post-flight matters: gimbal camera maintenance covers acclimatising sealed and the condensation cycling that corrodes connectors over a season.
Deciding Whether to Fly
Two separate questions, and conflating them is how surveys get wasted.
- Will the payload survive? Check ingress rating and temperature envelope against the conditions.
- Will the imagery be valid? Apply the derate and ask whether the mission is still achievable at the reduced range.
- If the answer to the second is no, is a closer-standoff version of the mission worth flying instead?
- Record the conditions either way — they bound what the survey could see and protect the report.
For inspection work with a timing constraint — envelope surveys, roof moisture, turbine delamination — the honest answer is often to reschedule. Those surveys have narrow valid windows and a marginal-weather flight produces imagery that looks professional and diagnoses nothing.
For search and emergency work the calculation is different, because a degraded search still beats no search. Tighten the pattern, reduce standoff, and record the conditions so the coverage claim is honest.
Related reading
- From NETD to DRI
- Thermal Infrared Optics
- Why Thermal Cameras Live at 8–14 µm
- DRI Ranges Explained
- How Far Can a Drone Thermal Camera Detect a Person?
- Thermal Sensitivity (NETD) Explained
- IP Ratings on UAV Payloads
- Cold-Weather Thermal Drone Operations
- MV-2P — <40 mK at 130 g
- OP-90A — zoom, thermal and ranging
- LX-9B — long-range multi-sensor pod
- Thermal Drone Payloads for Search and Rescue
FAQ
Can thermal cameras see through fog?
Only marginally better than visible cameras in dense fog. LWIR penetrates light haze well — that is a genuine advantage — but thick fog scatters it strongly because the droplets are close in size to the wavelength, and range collapses. Plan on anything beyond a few hundred metres being unreliable in real fog, and do not let the “thermal sees through anything” reputation survive contact with your mission planning.
Does rain damage the payload?
A well-sealed payload survives rain within its ingress rating, but survival and usability are different questions. Water on the lens degrades the image and shifts thermal accuracy, because the film becomes an emitting surface between sensor and scene. Check the rating for the exposure, then separately ask whether the imagery will still be valid — and note that IP ratings cover static immersion rather than rain at flight speed.
Is there any weather that improves thermal performance?
Cold, dry, still air is excellent — high contrast and low attenuation, which is why winter search and envelope work read so well. Roof moisture surveys also improve shortly after rain, once the surface has dried, because evaporative cooling deepens the contrast between wet and dry sections. In both cases the limiting factor moves from the atmosphere to batteries and crew endurance.
Questions about the technology? Talk to our engineers — we reply within 2 business days.