The orange-and-purple picture that a thermal camera makes of a house or a person is not a photograph of colored light. The camera is blind to color; it images warmth — infrared radiation your eyes cannot see, the same invisible emission responsible for why two same-colored cars heat up differently in the sun — and then paints the resulting numbers with a lookup table. Iron reds for the hottest targets, fading through oranges to purple-blue for the coldest. Nothing in the scene was orange. The palette is an editorial decision made by engineers, and the same thermal data can honestly wear grayscale, ironbow, rainbow or 'arctic' with equal right to be called accurate. Understanding this changes how you watch every thermal video: you are not seeing heat's true colors, because heat has none. You are seeing a chart that has been dressed as a photograph.
From warmth to numbers to colors
Step by step, the pipeline runs like this. Each sensor element reads incoming infrared radiation and converts it to a signal; calibration turns that signal into an estimated temperature for whatever surface lies in that direction; and then a colorization step maps every temperature to a color so that human eyes — superb at comparing hues, poor at comparing brightness levels — can read the pattern instantly. Textbook treatments of infrared thermal imaging describe exactly this chain, from radiance through calibration to the final displayed thermogram. By the time color is assigned, the physics is finished. The camera is not recording that the wall is red; it is deciding that 30 degrees shall be called red.
Two consequences follow immediately. Different cameras — or the same camera in different modes — can dress one scene in entirely different costumes while every reading stays honest. And the 'photograph' you are watching is closer kin to a shaded relief map than to a picture: each color is a legend entry, not a property of light that reached the sensor as that color.
Palettes change what you notice
The choice of palette is not cosmetic. Every palette compresses a range of temperatures into a small set of visible steps, and the placement of those steps determines which differences shout and which whisper. Grayscale preserves smooth gradients and is often preferred for reading fine spatial detail. High-contrast palettes like ironbow make hotspots pop out fast, which is why building inspectors favor them for finding missing insulation. Rainbow palettes look vivid and information-rich, but they have a well-documented cost: they introduce artificial boundaries — sharp jumps where the data is actually smooth — and can imply structure that does not exist, a problem visualization researchers have criticized for decades. Recent experimental work on infrared imaging perception confirms the practical stakes: palette choice measurably affects how accurately people interpret thermal images.
The infrared world leaks into ordinary life more often than you might expect, always invisible until a device translates it. The flashing light in a TV remote is near-infrared that a camera happily records; a thermal camera is the same trick pushed to longer wavelengths and dressed in a palette. Both devices are choosing, on your behalf, what invisible radiation should look like.
The palette question matters most where decisions follow the picture. An inspector scanning a wall for missing insulation is not looking for 'the orange parts' — they are looking for a temperature step, and the palette decides whether that step lands on a visible color boundary or hides between two hues that read almost the same. Wildlife filmmakers face a related choice: a grayscale night scene reads as documentary sobriety, while the same frames in ironbow read as spectacle, with no change in the underlying data. Even the screening cameras that appeared in public spaces dress their output in palettes chosen for glanceability, though a camera's temperature estimate is shaped by distance, angle and surface as much as by physiology — a boundary worth remembering whenever a colorful thermal image is presented as a measurement of a person.
Reading a thermogram like a chart, not a photo
A few habits make thermal images harder to fool you. Read the temperature scale before the picture — the bar that ties colors to numbers is the actual content, and everything else is presentation. Remember that the palette is arbitrary: the same wall can legitimately be white-hot in one image and violet in another, and two different cameras may dress identical temperatures differently. Remember that smooth data can be made to look banded by the palette, and banded data can be made to look smooth. And remember that surface readings have their own physics — shiny metal and glass reflect infrared, so a thermal image of a window can show the photographer rather than the room, and a polished saucepan can mirror the ceiling. None of this makes thermograms untrustworthy; it makes them interpretive documents, closer to diagrams than to photographs.
The skill of noticing where a color scale's steps land is the same one Hue Sorting sharpens — ordering neighboring hues along a smooth ramp is exactly what a palette asks your eye to do. The next time a thermal image flashes past in a documentary, ask it one question first: who chose these colors, and what did they want to stand out?
Same temperatures left to right in every row — only the emphasis changes: where each palette places its sharp steps.