How Thermal Imaging Works: The Physics in Plain Language
Thermal imaging looks like magic from the outside — a camera that “sees” heat. The underlying physics is well-defined and worth understanding because the limits of the technology fall directly out of those physics. This guide walks through the long-wave infrared band, the detector technology, how surface emissivity affects readings, and how a temperature image gets built from invisible radiation. Material draws on energy.gov, ASHRAE, and InterNACHI references current as of 2026.
Everything Above Absolute Zero Emits Radiation
Any object at any temperature above absolute zero (-273.15°C) emits electromagnetic radiation. The wavelengths and intensity of that radiation are determined by the object’s temperature, described by Planck’s law and summarized in the everyday observation that hotter objects glow at shorter wavelengths. A red-hot stove burner emits visible red light at around 1000 K. The sun at roughly 5800 K emits across the visible spectrum, which is why sunlight appears white.
Objects at terrestrial temperatures — buildings, people, electrical equipment, the ground — emit primarily in the long-wave infrared range, roughly 8 to 14 microns. Human vision tops out at around 0.7 microns (red light). The radiation buildings emit is invisible to human eyes but real and measurable.
The 8 to 14 Micron Window
The 8-to-14-micron band is called the long-wave infrared (LWIR) window. It is useful because Earth’s atmosphere is largely transparent in this range — water vapor, carbon dioxide, and other atmospheric gases absorb relatively little here compared to other infrared bands. That means radiation from a wall 20 feet away reaches a camera lens without being attenuated to noise.
Thermal cameras designed for building, electrical, and industrial inspection operate in this LWIR window. Mid-wave infrared (3 to 5 microns) is used in some specialized applications where target temperatures are higher; near-infrared (around 1 micron) is used in night-vision systems but does not measure thermal emission. Each band has different physics and different detector requirements.
Microbolometer Detector Arrays
The sensor inside a modern thermal imaging camera is a microbolometer detector array. Each pixel in the array is a tiny resistor whose electrical resistance changes with temperature. Incoming infrared radiation warms the pixel slightly, the resistance changes, and the readout electronics convert that change into a digital value. Doing this across an array of 80×60, 320×240, or 640×480 pixels produces an image.
Microbolometers are uncooled — they operate at room temperature without the cryogenic cooling required by some scientific infrared detectors. That uncooled operation is what made thermal imaging practical and affordable for field use. The trade-off is sensitivity: uncooled microbolometers cannot detect temperature differences as small as cooled detectors can, but the difference is fine for most building and electrical applications.
From Radiation to Temperature
The camera processor uses the Stefan-Boltzmann law and the assumed emissivity of the target surface to convert detected radiation into a temperature reading. The math is simple in concept: the radiation flux at the detector is proportional to the surface temperature to the fourth power, multiplied by emissivity, plus reflected ambient radiation. Solving for temperature gives the value the camera displays.
This is why emissivity settings matter. If the camera assumes ε = 0.95 (typical for painted drywall) and the target is actually bare polished aluminum at ε = 0.10, the temperature reading will be wildly wrong because the math assumed a surface that emits ten times more strongly than it actually does. Working thermographers learn to recognize low-emissivity materials and adjust accordingly.
What Determines Emissivity
Emissivity is a property of the surface, not the bulk material. Polished metals have low emissivity (0.05 to 0.20) because their conductive electrons reflect more than they emit. Painted, oxidized, or anodized metals have high emissivity (0.85 to 0.95) because the surface layer behaves more like a dielectric. Most building materials — painted wood, painted drywall, brick, concrete, asphalt shingle — have emissivity around 0.90 to 0.95, which is why default camera settings work well on residential interiors and exteriors.
Glass and water have emissivity around 0.92 to 0.96 in the LWIR band, so they read accurately. But glass is opaque to LWIR even though it is transparent to visible light, which means a thermal camera does not see through windows; it sees the temperature of the glass surface. The same applies to thin plastic films and most ceramic glazes. Sibling reading: how does infrared camera work.
Reflections and Ambient Radiation
A surface in front of a thermal camera radiates from its own temperature and also reflects ambient infrared radiation from objects in front of it. A polished metal panel mostly reflects, so it shows the thermal signature of whatever is in front of it — including the warm body of the operator holding the camera. This is the source of the “hot spot that moves when I move” effect that novices misinterpret as a real anomaly.
For accurate temperature reading of a low-emissivity surface, thermographers add reflective tape with known high emissivity, allow it to equilibrate to the surface temperature, and measure the tape rather than the bare metal. This is standard practice for measuring electrical bus connections or similar polished components.
How a Thermal Image Gets Built
The detector array reads out continuously, typically at 9 to 60 frames per second depending on the camera tier. Each frame is a grid of temperature values. The processor maps that grid to a color palette — iron, rainbow, grayscale, lava — assigning a color to each temperature within a dynamic range. The result is a false-color image where color encodes temperature.
The dynamic range is adjustable. Setting a tight range around the temperatures present in the scene maximizes contrast and makes subtle anomalies visible. Setting a wide range compresses contrast and hides detail. Most cameras offer auto-ranging and manual ranging; analytical work usually uses manual ranging so successive images can be compared on the same scale.
Spatial Resolution and the Detector Array
The detector resolution determines how much spatial detail the image carries. An 80×60 detector produces 4,800 temperature data points per frame; a 320×240 detector produces 76,800; a 640×480 detector produces 307,200. More pixels resolve smaller targets at a given distance. Consumer-tier cameras typically run 80×60 to 320×240; professional handhelds 320×240 to 640×480; specialized scientific units go higher.
The phrase “HD thermal” applied to 1024×768 cameras is industry shorthand, not a true HD-video equivalent. Software upscale features like FLIR’s MSX, Seek’s SeekFusion, or similar overlay visible-light edges onto the thermal image to make it look sharper but do not increase the underlying thermal resolution. Hub reading: home inspection tools hub.
Thermal Sensitivity (NETD)
Noise-equivalent temperature difference (NETD) measures how small a temperature difference the camera can distinguish from sensor noise. Typical specs run 30 to 100 millikelvin (0.03 to 0.10°C). Lower NETD means the camera can resolve subtler temperature gradients, which matters for building envelope work where stud bay vs. cavity differences might be only 1 or 2°C.
NETD specifications are often given at 30°C ambient and may degrade in cold conditions. A camera spec’d at 50 mK at 30°C may show 80 to 100 mK in attic temperatures around 0°C. Working inspectors learn the cold-weather behavior of their particular unit.
Atmospheric Effects at Distance
The LWIR window is mostly transparent over building-inspection distances (under 100 feet). Over longer distances, water vapor and CO2 absorb some of the signal, requiring atmospheric compensation in the camera or in post-processing. For industrial use cases measuring transmission lines or large structures at hundreds of feet, atmospheric correction matters; for residential and most commercial inspection, it does not.
Rain, snow, and fog attenuate the signal more strongly. Imaging a wet exterior wall produces results affected by evaporative cooling and by the absorption of LWIR by liquid water in the path. Working thermographers note weather conditions on every inspection report.
Why All of This Matters for Inspection
Understanding the physics changes how the image is read. A “hot spot” on a metal panel is probably a reflection. A “cold streak” on a sunlit wall is probably solar gain mismatch. A “moisture pattern” on a ceiling that has not been wet recently might just be a cooler air mass. The image shows what the surface is emitting; the inspector translates that into what is going on behind.
This is why training and certification programs exist for thermography. The hardware is increasingly accessible at consumer prices. The interpretation skill that turns the image into a correct diagnosis is what makes the tool worth more than a curiosity.
Common Misconceptions About What Thermal Imaging Sees
The most persistent misconception is that thermal imaging “sees heat” the way regular cameras see visible light. The two share the surface-imaging principle but differ on what they detect. Visible cameras detect reflected light from external sources (sun, room lighting). Thermal cameras detect radiation emitted by the surfaces themselves. A dark room is invisible to a regular camera but fully visible to a thermal camera as long as objects in the room are at temperatures different enough to distinguish.
The second misconception is that thermal cameras “see through” walls. They do not. They see surface temperatures, which can reveal patterns of heat flow behind the surface but only when those patterns produce surface temperature differentials. A wall with perfectly uniform insulation and no air leakage shows uniform color regardless of what is in the cavity.
The third misconception is that thermal images give absolute temperatures with high precision. Consumer-grade cameras specify ±5°C accuracy; professional cameras ±2°C. For comparing one part of a scene to another, this is fine. For absolute measurement, the spec matters.
Comparing Thermal to Other Inspection Technologies
Thermal imaging complements rather than replaces other inspection methods. Moisture meters give point readings at higher precision than thermal estimates of evaporation; the combination of thermal screening plus moisture confirmation is more useful than either alone. Borescopes inspect cavities directly without thermal inference, useful when thermal patterns are ambiguous and a direct look would resolve them.
Blower doors quantify air leakage by measured airflow at known pressure differential; thermal imaging visualizes where that air moves but does not quantify the volume. The combination is standard practice in energy audits. Microwave-band moisture detection penetrates deeper into substrates than evaporative thermal signals but costs significantly more. Each tool occupies a niche; thermal imaging covers wide-area screening fast and points to where to look more carefully.
How the Industry Talks About These Specs
Manufacturer marketing emphasizes whichever specs put the product in the best light. Detector resolution gets prominent placement because consumers understand pixels. NETD often appears in fine print because lower numbers are better and the term is unfamiliar. Refresh rate matters for video applications but is less relevant for inspection-style still capture; export-controlled categories cap consumer units at 9 Hz.
Accuracy specs are typically given at a controlled ambient (often 30°C) and across a specified range (often -20°C to +120°C for inspection cameras). Performance outside that range is extrapolated and less reliable. Reading the spec sheet carefully, rather than just the headline number, matters when choosing a camera for serious work. Sibling reading: how does thermal imaging work on the human body covers another application angle.
References
- Thermographic Inspections — U.S. Department of Energy
- Infrared Thermal Imaging Inspection — InterNACHI
- Standards and Guidelines — ASHRAE
- Professional Home Energy Assessments — U.S. Department of Energy
Thermal imaging cameras
Infrared cameras reveal hidden moisture, missing insulation, and air leaks. Phone-attachment models are the budget entry point; standalone units have higher resolution.
| Product | Why | Buy |
|---|---|---|
 FLIR ONE Pro (phone) | Plugs into iPhone/Android; inspector favorite. | Amazon — $349.00 |
 Topdon TC001 | High-res phone module at a low price. | Amazon — $199.99 |
 FLIR C5 Compact | Standalone pocket camera with Wi-Fi. | Amazon — $449.00 |