Point a thermal camera at a polished copper heating pipe carrying water at 60°C. Depending on conditions, the screen will show 23°C. The device isn't broken, it isn't poorly calibrated, and there's no manufacturing defect.
This is the most common mistake in thermography, and it comes down to a single concept.
What the sensor actually measures
A microbolometer doesn't measure a temperature. It measures an amount of infrared radiation reaching it. To infer a temperature from that, it applies the Stefan-Boltzmann law, which requires knowing one property of the observed surface: its emissivity.
Emissivity ranges from 0 to 1. It describes a surface's ability to emit its own radiation rather than reflect that of other objects. A perfect black body would be 1. In real life:
| Surface | Emissivity | Reliable measurement? |
|---|---|---|
| Matte paint, plaster, brick | 0.90 to 0.95 | yes |
| Wood, paper, fabric | 0.85 to 0.95 | yes |
| Human skin | 0.98 | yes |
| Glass | 0.85 to 0.95 | yes, but reflective at an angle |
| Oxidized steel | 0.70 to 0.80 | correction needed |
| Anodized aluminum | 0.55 to 0.80 | correction needed |
| Polished stainless steel | 0.10 to 0.20 | very difficult |
| Polished copper, polished aluminum | 0.03 to 0.10 | nearly impossible |
Polished copper at 0.05 emits only 5% of the expected radiation. The remaining 95% the camera sees is the room's own reflection.
The second trap: reflected temperature
If a surface reflects, you need to know what it reflects. That's the role of reflected temperature, sometimes called background temperature or apparent ambient temperature.
On a copper pipe in a 20°C boiler room, the camera mostly picks up the reflection of the 20°C walls. It therefore displays a value close to 20°C, regardless of the water temperature inside.
Setting emissivity without setting reflected temperature isn't enough: the two corrections work together. That's precisely why some manufacturers automate both in a single step.
How to get an accurate reading on metal
The tape method. Stick a piece of matte insulating tape on the surface, wait a minute for it to reach thermal equilibrium, and measure the tape with an emissivity of 0.95. You're then measuring the tape, which is at the pipe's temperature. It's the simplest and most reliable method.
The paint method. A dab of matte black paint, a marker, or even some white correction fluid works on the same principle.
The reference-point method. Take a contact reading with a thermocouple, then adjust the camera's emissivity until it displays the same value. The FLIR TG267 is the only device in our catalog shipped with a type K thermocouple, which makes this method directly usable.
The automatic method. testo supplies reference stickers, the ε-Marker, with a known emissivity. The device determines the surface's emissivity and the reflected temperature on its own. Three of these stickers ship standard with the 871s.
What the datasheets reveal
Of the 101 devices in our catalog, 36 document a continuous emissivity setting, two offer preset values, and 58 document nothing at all.
For reflected temperature, it's worse: 24 devices document it, 72 stay silent.
Be careful with interpretation, though. "Undocumented" doesn't mean "absent". We checked the case of HIKMICRO's B series: neither the B20's nor the B20S's datasheet mentions emissivity or reflected-temperature correction. The user manual, however, describes both settings, along with distance and humidity correction.
In other words, the manufacturer highlights a heavily interpolated resolution while staying silent about the functions that determine measurement accuracy. That's a communication choice that says a lot about the buyer it's targeting.
The two features that actually change everything
ε-Assist, at testo. The device automatically determines emissivity and reflected temperature using a reference sticker. The two settings that skew measurements the most, solved with no intervention needed. No other manufacturer in our database offers this.
The full correction chain. Emissivity, reflected temperature, ambient temperature, humidity, and distance. Three manufacturers document it in full: testo, FOTRIC, and Fluke. HIKMICRO implements it on its B series but only publishes it in the manual.
What to check before buying
A freely adjustable emissivity value, not just a list of preset materials. On a circuit board or an industrial installation, you'll encounter surfaces that appear on no list.
A reflected-temperature setting. Without it, emissivity correction remains incomplete as soon as you go below 0.7.
The ability to correct after the fact. This is where full radiometry becomes decisive: if every pixel retains its value, you can reopen a poorly set image and redo the calculation at your desk. Otherwise, the reading is lost and you have to go back on site.
The comparator shows these three criteria for every device, and flags the ones whose manufacturer publishes nothing on the topic.
The rule of thumb to remember
On matte, non-metallic surfaces, which make up most building diagnostics work, a camera set to 0.95 gives accurate results with no extra effort.
The moment bare metal, angled glass, or a shiny surface appears, don't trust any number on screen until you've stuck a piece of matte tape on it.
All values cited come from manufacturer datasheets and the sources listed on each product page. No data is estimated.