Thermal cameras: what a temperature difference can and cannot prove

Thermal imaging produces the most persuasive picture in residential surveying and the least self-explanatory one. A blue patch on a wall looks like a finding. It is not. It is a surface a degree or two cooler than its surroundings, and there are at least five ordinary reasons for that, only one of which is a building defect. The surveyors who get into trouble with a thermal camera are not the ones who cannot operate it; they are the ones who write the picture up as if it were a diagnosis.

What it measures

A thermal camera does not see heat. It measures the long-wave infrared radiation leaving a surface, applies an assumed emissivity — how efficiently that surface radiates compared with a perfect black body — subtracts an assumed reflected temperature, and presents the result as a colour map of apparent surface temperature. Three things follow.

First, it is a surface instrument. The wall two centimetres in is invisible to it. What you see is the outer skin, and the only way the interior shows is by changing the temperature of that skin — which is why conditions matter so much.

Second, the number on screen is only as good as the two assumptions. Masonry, plaster, timber, matt paint and most building materials sit at an emissivity of roughly 0.90–0.95, which is why the default setting works often enough to be dangerous. Glass, glazed tiles, polished metal, foil-faced insulation, stainless flues and gloss paint do not; they act as infrared mirrors and show you a reflection of yourself, the room or the sky. A radiator “cold spot” that follows you across the wall is your own reflection.

Third, the camera’s resolution and thermal sensitivity set what is physically resolvable. Resolution is the detector array: 160 × 120 is 19,200 measurement points, 256 × 192 is 49,152, 320 × 240 is 76,800. Sensitivity — NETD, quoted in millikelvin — is the smallest temperature difference the detector can distinguish; under 50 mK is good, under 40 mK is very good. Marketing figures such as FLIR’s MSX or Testo’s SuperResolution are an overlay of visible-light edges or an interpolation; they make the image easier to read but do not add thermal information.

Where it misleads

• Evaporative cooling looks like damp — and so does everything else cold. A damp patch is cooler because water evaporating from it takes heat with it. A thermal bridge at a lintel, a missing piece of insulation, an air leak around a window, and the shadow of a tree that moved an hour ago are also cooler. The pattern helps — damp tends to be irregular with soft edges and to sit where water would go — but the camera cannot tell them apart. A moisture meter profile can.
• Yesterday’s sun is still on the wall. Masonry stores heat for hours. An elevation that had sun on it in the afternoon will show a mottled, meaningless pattern until the small hours.
• Not enough temperature difference. Building-envelope thermography needs the inside and outside to differ by at least 10 °C, and to have done so for several hours, so that heat is flowing steadily through the fabric. In a British July, or an unheated void property, there is often nothing to see, and an image taken anyway proves nothing either way.
• Wind and rain. Wind strips the surface temperature difference off an external wall; rain wets everything uniformly and masks the one patch that matters.
• The palette. Auto-ranging rescales the colours to whatever is in frame. The same 1 °C difference can look dramatic or invisible depending on what else is in shot. Lock the span and level before you compare two images.
• Indoors, furniture and heating. A sofa against a wall leaves a warm shadow for an hour after it is moved. A radiator or a sunlit window dominates the frame and compresses everything else into one colour.

Using it so the result holds

Treat the camera as a screening tool that tells you where to put the moisture meter, the borescope or your hand. The 2022 RICS, Historic England and PCA joint position statement on moisture investigation describes it in exactly those terms, and that is the standard a report will be judged against.

• Record the conditions. Inside and outside air temperature, relative humidity, wind, cloud, rain in the previous 24 hours, when the sun was last on the elevation, and whether the heating was on and for how long. Without these a thermal image is uninterpretable later.
• Set emissivity and reflected temperature for the material in frame, or state that you did not and that the temperatures are therefore indicative only.
• Shoot a visible-light reference of the same view, from the same position, every time. The camera’s own visual channel is fine. An anomaly with no visible reference cannot be located later.
• Confirm with a second method before it becomes a finding. A cool patch becomes “an area of elevated comparative moisture readings” only after the meter says so; a cold band at a lintel becomes a thermal bridge only when its position matches the construction.
• Scope. A thermal survey of a building envelope to BS EN 13187 — the standard for qualitative detection of thermal irregularities — is a separate instruction with its own conditions, usually at night. Pointing a camera at a wall during a Level 2 inspection is a screening observation and should be described as one.

What to buy

Model	What it is	Indicative UK price (inc VAT, Oct 2026)	Verdict
HIKMICRO Pocket2	256 × 192 radiometric, NETD under 40 mK, 25 Hz, visual camera and fusion, Wi-Fi, free analysis software	About £500	The sensible surveyor’s camera. Enough resolution to locate a cold bridge on a two-storey elevation, enough sensitivity to see evaporative cooling, and it fits in a jacket.
HIKMICRO B20	256 × 192 in a pistol-grip body, visual camera, larger screen	About £510–£610	The same sensor with a trigger grip. Buy it over the Pocket2 if you prefer a camera-shaped camera.
Teledyne FLIR C5	160 × 120 radiometric pocket camera, MSX overlay, cloud upload	About £635–£650	Overpriced for the sensor. It will tell you where to put the meter, which is all a 160 × 120 array can do, but the Pocket2 does that better for less.
Testo 872s	320 × 240 native, about 50–60 mK, building modes including a humidity traffic-light with a Bluetooth probe	About £2,450	The step up if you sell thermography as a service in its own right. The humidity probe integration is genuinely useful for condensation work.
Teledyne FLIR E8 Pro	320 × 240, under 40 mK, MSX, swappable batteries, FLIR’s cloud reporting	US list $3,299; about £3,000–£3,500	Professional-grade and built for all-day use. Only worth it over the Testo if you are already in FLIR’s software.
Testo 883-1	320 × 240 with SuperResolution to 640 × 480, under 40 mK, manual focus, interchangeable lenses	About £4,200 (Testo UK)	Thermographer’s kit. Not a Level 2 purchase.
Phone-mounted (HIKMICRO Mini2, FLIR One Edge Pro)	256 × 192 or 160 × 120 dongles	About £250–£500	Useful as a first camera. The phone, the cable and the app are the weak points on a cold roof.

The one to buy for most residential work is the HIKMICRO Pocket2. It is radiometric, so the images can be reanalysed later with the span, level and emissivity changed, which a cheap non-radiometric camera cannot do; it has the sensitivity to show the differences that matter; and at about £500 it costs what the FLIR C5 costs with four times the pixels.

What the extra money buys above that is resolution — a 320 × 240 array resolves a smaller patch from further away, which matters on a three-storey elevation or a long roof slope — and build quality for daily use. It does not buy you a diagnosis.

Not worth paying for: “super-resolution” as a headline figure; cloud subscriptions for a camera you use twice a month; a 640 × 480 camera for residential condition work; and any camera that will not save radiometric images, because a pretty JPEG you cannot reinterpret is a liability when it is challenged.

Training is worth more than the next camera up. A Category 1 thermography course (BINDT/UKTA or equivalent, typically four days) is the point at which the images stop being pictures and start being measurements.

Prices move; check the live price before buying.

Recording it in the report

Weak wording: “Thermal imaging identified damp to the rear wall of the kitchen.”

Wording that holds: “A thermal image of the rear kitchen wall (HIKMICRO Pocket2, emissivity 0.95, internal 19 °C, external 6 °C, overcast, heating on since 06:00) showed an irregular area of lower surface temperature about 400 mm wide at low level, left of the sink. Pin-mode moisture readings across the same area were elevated relative to a control on the adjacent wall. The pattern is consistent with moisture in the plaster at that location; the camera does not identify the source. See the moisture readings and the recommendation at D5.”

Instrument, conditions, settings, pattern, the confirming method and the inference — kept separate — are what turn an image into evidence.

References

• BS EN 13187:1999, Thermal performance of buildings — Qualitative detection of thermal irregularities in building envelopes — Infrared method, BSI.
• RICS, Historic England and PCA, Investigation of moisture and its effects in traditional buildings — joint position statement, 2022.
• BRE, Thermal imaging of buildings (BRE guidance on survey conditions and interpretation).
• BINDT/UKTA, Category 1 thermography training syllabus.
• RICS, Home Survey Standard, 1st edition, 2019 (effective March 2021).