When a Thermal Camera Sees a Reflection
A shiny tray and a matte patch can share a temperature yet look nothing alike. The camera is responding to infrared radiation, some of it sent by the room.

In this story 4 sections
There is an especially good way to appear in a self-portrait without turning a camera toward yourself. Point a thermal imager at a polished piece of metal and move to one side. A bright patch may shift toward the edge as you go. If it looks surprisingly like an upside-down version of a sleeve and shoulder, the metal has given you some evidence about the picture: the warm part you are seeing may be standing on the other side of the camera. A reflected outline can be startling if you expected every shape in the view to belong to the object in front of the lens.
In ordinary light, we accept from experience that some objects show us parts of the room behind us. A pot lid can contain a whole kitchen for the length of a glance. Infrared imaging deals with a different part of the spectrum, so the visual estimate of which object might reflect can be incomplete. What carries over is the distinction between light coming from a surface and light sent toward it from somewhere else and reflected into the camera. Once you have a way to separate those possibilities, the false-color scene becomes considerably easier to read.
The orange is an assignment
A thermal camera receives infrared radiation through its optics. Its software can assign colors according to the signal or the temperatures estimated from it under specified assumptions. That is why orange is not ordinarily a property of the object itself. If you select a grayscale palette you can often tell much the same story in white, gray and black. You can also change which temperatures correspond to the extremes of the color scale. A vivid small patch in a tightly limited scale can look more important than the same patch shown as part of a wider range.
Consider a view of an indoor wall saved as a reference. Later you frame it again, but a sunny area near the window occupies part of the new image. If the software chooses its scale automatically it can give the wall different colors simply because the hotter area now has to fit on the same scale. Two things moved at once: what the camera sees and how the camera decides to show it. Locking the scale for the comparison lets you set one of those factors aside. A pair of saved pictures is easier to review when each keeps a readable legend as well as the visible framing.
The distinction matters more when someone sees only a cropped attachment, with no horizon and no scale. Before saying where the warmest item is, see whether the color scale was preserved and whether the image uses a mode that blends a normal photo with infrared data. Those helpful little edges may locate a panel or a fixture without having been measured by the thermal detector. They can make a rough boundary look more precise than the underlying infrared reading.
A surface can contribute less of its own
Emissivity describes how well a surface emits thermal radiation compared with an ideal emitter at the same temperature. A comparatively high-emissivity surface can give the camera a strong contribution from the surface itself. A low-emissivity opaque one can send much more of what it receives back toward the room. Both can have the same physical temperature. For an imager that takes its settings or assumed values as given, the readings on the two surfaces can differ considerably.
Polished metal often supplies the demonstration, but the form of the surface matters as much as the noun used to name it. A very polished plate and a dull painted plate can behave differently even when the metal underneath is identical. A kitchen cabinet handle may contain a glossy stretch, worn patches, and places where a hand touched it a moment ago. These are three reasons for apparent differences in a small area, only the last of which immediately implies a recent change in temperature. It is possible to notice the difference without being able to choose a cause from one image.
Compare with a simple thought experiment. Imagine looking at a reflective empty tray indoors when there is a warm object off to one side. Parts of the tray could look hotter than the rest. If the reflected source were instead a relatively cool part of the room, those same parts might look unusually cool. A reflected signal does not give an error of one consistent sign. Treating shiny things as always warmer in a thermogram would be a mistake nearly as quickly as assuming they always look colder.
A quiet scene you can check twice
A clean, room-temperature tray is useful for learning the difference because you can do the test in a safe setting with no heated, powered or moving equipment. Place an ordinary matte sticker on a small part of the tray and let everything settle in the same room. Look first at the tray and the sticker from an angle, then shift the camera to another position. Keep your recording scale stable. If a shape slides over the shiny area with you or moves off the edge while the sticker area stays comparatively stable, you have some evidence of a reflection. The whole tray did not have to change temperature as fast as you moved.
No part of that experiment assigns a known emissivity to an arbitrary sticker or makes a reference instrument out of the camera. FLIR describes a more controlled version of the method that uses a suitable high-emissivity surface in contact with a lower-emissivity material. The reading is of the surface seen by the detector; for a useful comparison the reference patch needs time to reach a similar temperature, and must be large enough relative to the measurement area. This is an exercise about what the image contains, not certification of an accurate instrument.
There are several good reasons it might be hard to see a reflection. The room may be close to one temperature everywhere. An area may be too small or too far away and get averaged with the surrounding picture. A deeply curved object may send an easily recognized shape apart into smaller patches. Try lowering the camera rather than moving the whole table. Now the background the object sees changes without rearranging the test object or moving it to a draft. The visible consequences are easier to interpret with fewer moving parts.
From a room to an inspection picture
A thermal view can be a great way to locate differences across a wide area. The trouble comes when a difference is treated as a complete explanation. A cool-looking region of one wall can have several causes. Maybe a change in insulation or draft has affected the surface temperature. A seam may use a different finish. Reflected radiation that matters on a glossy panel could be largely absent on an adjacent matte one. To establish which explanation is appropriate requires knowledge of the material, the surroundings and the conditions at the time. An image showing a pattern is many steps away from a recommendation to open a wall.
Make a small field card if you are comparing pictures for an ordinary observation. Record the date, approximate distance, angle, surface material and whether the scale was automatic. Record whichever nearby warm or cool objects might have been reflected. A separate view in visible light can preserve that context. If you have a comparison frame from a second angle, write what moved relative to the physical object rather than trying to describe the color alone. A useful note might point out that a pale comma followed the camera past the handle while a square patch remained attached to the fastening. The later reader can then see why you distinguished those two effects.
The material of an object can vary within its outline, too. Suppose the two sides of a hinge have similar functions, but one side was painted over and the other is still bright. It is easy to plan a neat left-to-right comparison and entirely miss the difference in finish. Checking a visually uninteresting detail before trying to explain a dramatic temperature pattern is often the more productive order of work. If the readings will matter for a repair, diagnosis or inspection, they deserve an appropriate method and independent confirmation. The camera is one instrument in that decision.
That distinction also matters when saving a sequence of images. If an apparent hot spot occupies just one detector pixel, it can look larger after the viewer’s software interpolates the image for a display. Enlargement helps you find the pixel again; it does not make the measurement more local. Save one image that shows the whole target in context and another at a useful working distance. Compare the smaller region with the pixel grid or the manufacturer’s specified measurement spot instead of trying to count pixels in a printed page.
Return to the bright patch in the tray after looking at something matte. Notice how quickly a face or sleeve in a reflection comes to feel as obvious as an ordinary mirror image. That recognition gives you a new habit: when a color calls for attention, ask first which surface is sending the signal into the lens and what else that surface can see. Sometimes the most useful thing an image tells you is that the room is bigger than the frame.


