How Does Thermal Imaging Work: Physics and Sensor Guide
“How does thermal imaging work” gets asked at every level — by homeowners considering a thermal camera purchase, by inspectors using the technology daily, and by anyone watching thermal footage in news reports or wildlife documentaries. The answer involves three layers: the physics of thermal radiation, the sensor technology that detects it, and the image processing that translates measurement into visible output. This guide walks through each layer at a level appropriate for non-physicists who want a real understanding rather than marketing copy. The framework draws on Department of Energy thermography publications and InterNACHI training materials current as of 2026.
The Physics: Stefan-Boltzmann Radiation
Every object with temperature above absolute zero (-273.15°C, 0 Kelvin) emits electromagnetic radiation. The Stefan-Boltzmann law describes the total emitted power: P = εσT⁴, where ε is emissivity (a surface property from 0.0 to 1.0), σ is the Stefan-Boltzmann constant, and T is absolute temperature in Kelvin. The fourth-power dependence means small temperature changes produce significant radiation changes — a 10% temperature increase produces ~46% more emitted radiation.
The spectrum of emitted radiation also depends on temperature, per Planck’s law. Cool objects emit primarily in long-wave infrared (8-14 microns). Warmer objects emit in shorter wavelengths. Very hot objects (above ~500°C) begin emitting visible light, which is why hot metal glows red. Building-temperature objects emit almost entirely in LWIR, which is the band thermal cameras for residential use are designed to detect.
What Emissivity Actually Means
Emissivity (ε in the Stefan-Boltzmann equation) is the efficiency with which a surface radiates compared to a perfect blackbody at the same temperature. A perfect blackbody (idealized) has ε = 1.0. Most natural and painted surfaces have ε between 0.85 and 0.95. Polished metals have ε as low as 0.05 — they reflect surrounding thermal radiation rather than emit their own.
For thermal imaging, emissivity matters because the camera measures emitted radiation and converts it to temperature using an assumed ε. If the actual surface ε differs from the assumed value, the calculated temperature is wrong. Most cameras allow emissivity adjustment; setting it correctly is essential for accurate measurement. Most building surfaces work well with ε = 0.95.
The Sensor: How Microbolometers Detect Infrared
Modern thermal cameras for residential use employ uncooled microbolometer detectors. Each pixel is a microscopic resistor suspended on a thin membrane that absorbs incoming infrared radiation. The absorbed energy heats the resistor very slightly, changing its electrical resistance in proportion to the radiation absorbed. Readout electronics measure the resistance changes across the array, producing a per-pixel signal proportional to the infrared radiation intensity at that pixel.
“Uncooled” means the detector operates at ambient temperature rather than requiring cryogenic cooling. Older technology — cooled InSb and MCT detectors — required cooling to ~80 Kelvin for sensitivity, which made cameras large and expensive. Microbolometers operate at room temperature, enabling handheld and battery-powered cameras.
The Optics: Why Germanium Lenses
Thermal cameras need lenses that transmit infrared. Normal glass transmits visible light but absorbs long-wave infrared, making it useless for thermal imaging. Germanium glass transmits LWIR efficiently and is the standard lens material for thermal cameras. Some cameras use chalcogenide glasses or specialty plastics as alternatives.
The need for germanium optics is why thermal cameras are expensive relative to visible-light cameras of similar resolution. Germanium is rare and difficult to process into precision optics. This is also why thermal cameras cannot see through windows — ordinary glass blocks LWIR.
The Signal Processing Chain
Raw microbolometer signals go through several processing stages before becoming visible images. Non-uniformity correction (NUC) compensates for variation between individual detector pixels. Temperature calibration translates resistance changes to apparent temperatures using factory calibration curves. Gain control adjusts the displayed temperature range to fit the visible palette range. Palette mapping converts temperature values to color or grayscale values for display.
Many cameras periodically pause for internal NUC by closing a shutter and using it as a known reference. This is the “click” sound thermal cameras make every minute or so. Without periodic NUC, drift accumulates and produces measurement errors.
Pseudo-Color Mapping Explained
Thermal images use “pseudo-color” — color that does not represent visible color of the scene but represents measured temperature. The translation is arbitrary and interpretive. The iron palette puts cold at black/blue/purple and hot at white/yellow; rainbow puts cold at blue and hot at red; grayscale puts cold at black and hot at white. The choice of palette affects how the image looks but not the underlying measurement.
Different palettes emphasize different features. Iron emphasizes hot spots. Rainbow emphasizes intermediate temperatures. Grayscale produces structural clarity without color distraction. Arctic emphasizes cool features. Inspectors typically specify the palette used in reports so readers interpret images correctly.
The Wavelength Choice
Residential thermal cameras operate in LWIR because that’s where room-temperature objects emit most strongly. The atmosphere is also relatively transparent to LWIR, which means imaging works through air over typical inspection distances. Other wavelength bands serve specialized applications:
MWIR (3-5 microns) is used for hot industrial targets where emission peaks at shorter wavelengths. SWIR (1-3 microns) and NIR (0.7-1.4 microns) detect reflected near-infrared rather than emitted thermal radiation — they require active illumination and serve remote sensing and certain night-vision applications. Residential thermal imaging is essentially synonymous with LWIR imaging.
Resolution and Spatial Detail
Thermal camera resolution is measured in detector pixels — each is an independent temperature measurement. Common resolutions in 2026: 80×60 (entry-level), 160×120 (mid-tier), 320×240 (inspector-grade), 640×480 (professional). Higher resolution produces finer spatial detail and more confident defect identification. The Department of Energy’s thermography guidance for building diagnostics generally recommends 320×240 minimum for inspector work.
Spatial resolution at the target depends on detector resolution, field of view, and distance to target. A 320×240 camera at 6 feet distance with a 25-degree field of view resolves approximately 0.3-inch features. The same camera at 20 feet resolves about 1-inch features. The pillar on home inspection tools covers the diagnostic toolkit.
Thermal Sensitivity (NETD)
Noise-Equivalent Temperature Difference (NETD) measures the smallest temperature difference the sensor can resolve above noise. Entry-level cameras spec around 100 millikelvin (mK). Mid-tier around 50 mK. Professional cameras at 25-30 mK. Lower NETD allows distinguishing subtler thermal patterns — useful in electrical inspection where early-stage faults produce small temperature elevations.
For building diagnostics with multi-degree thermal anomalies, NETD under 50 mK is generally sufficient. NETD is one of the specs that matters most for picking up subtle patterns; very low NETD (under 30 mK) is the difference between detecting an incipient electrical fault and missing it.
What Thermal Imaging Cannot Do
Thermal imaging detects surface temperatures. The camera does not see through walls — it sees the wall’s surface, and the temperature pattern at the surface provides clues about conditions behind. The cameras do not detect moisture directly; they detect thermal anomalies caused by moisture (evaporative cooling, conductive losses, capacitive properties).
Cameras do not produce useful images without temperature differential. Building diagnostics typically need 15-20°F interior-to-exterior differential. Shoulder seasons often produce flat images. Glass blocks LWIR, so cameras cannot see through windows. Solar loading on exterior surfaces creates patterns that can mask defects or produce false-positive artifacts. The cluster on how thermography works covers practical limitations.
How the Image Translates to Diagnostic Information
The thermal image is data. Pattern recognition converts data to diagnosis. Missing insulation produces discrete cold rectangles between studs on cold walls. Air leakage produces cold streamers radiating from penetrations during cold weather. Plumbing leaks produce localized cool patterns from evaporative cooling. Electrical hot spots produce warm patterns on panels and connections. Each diagnostic pattern is recognizable only with training in building science context.
The InterNACHI Infrared Certified credential covers pattern interpretation training. Uncredentialed thermal imaging produces images without diagnostic interpretation, which is why the credential matters when hiring inspection services.
Environmental Conditions for Useful Imaging
Effective thermal imaging requires conditions that produce thermal contrast. Interior-to-exterior temperature differential drives building thermal patterns. Solar loading can be helpful (for roof imaging) or harmful (creating exterior artifacts). Wind affects exterior imaging by changing convective heat transfer. Rain saturates surfaces and dampens patterns.
For Front Range conditions, winter mornings typically provide the best building imaging conditions — large interior-to-exterior differential without solar loading. Late-evening imaging avoids solar effects but may have lower differential. Plan imaging timing around the conditions that produce useful contrast.
From Physics to Practical Use
The complete chain: emitted infrared radiation governed by Stefan-Boltzmann and surface emissivity, captured through germanium optics, detected by uncooled microbolometer arrays, processed through NUC and calibration into pseudo-color images, interpreted through pattern recognition trained in building-science context. Each step matters for producing usable diagnostic results.
References
- Thermographic Inspections of Buildings — U.S. Department of Energy
- Infrared Certified Inspector Credential — InterNACHI
- Thermal Diagnostic Standards — ASHRAE
- Electrical Inspection Resources — Occupational Safety and Health Administration
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 — $329.00 |
 Topdon TC001 | High-res phone module at a low price. | Amazon — $199.99 |
 FLIR C5 Compact | Standalone pocket camera with Wi-Fi. | Amazon — $610.06 |