How can we observe if an object is a color we can't see?

The trick is translation, not seeing

An object doesn't have a single "color" floating inside it. It reflects or emits a mix of wavelengths, and "color" is what a particular eye does with that mix. Human eyes have three color-sensitive cone types, so wavelengths outside that range—ultraviolet beyond violet, infrared beyond red—simply produce no color sensation. To observe them, you need a device whose sensitivity extends into that range and then a way to display the result in colors you can perceive.

Infrared imaging works this way. A two-color infrared photothermal microscope excites molecules with infrared light and reads the resulting heat signal, producing microscopic images of two distinct infrared absorption bands and distinguishing two different chemical species inside live cells at sub-micrometer resolution [2]. The image you look at is a translation: the instrument measured infrared absorption, and the display assigned visible colors to represent it. Similarly, a near-infrared photothermal perception method converts subtle infrared differences into discriminative pseudo-color features so that they become distinguishable to a human viewer [5]. The colors in these images are labels for wavelength information, not the colors the object would appear to have.

If an animal can see it, let the animal tell you

For ultraviolet, one powerful approach is behavioral testing with an animal that naturally sees UV. Researchers built a five-channel LED display (red, green, blue, violet, and UV) and tested anemonefish, first measuring that the fish's UV cone peaks at about 386 nanometers—just below the human visible range—alongside three other cone types [1]. The fish were then trained to pick out a colored target pixel from grey distractor pixels of varying brightness. Because the fish used all four cones, it is tetrachromatic, meaning it has four color channels rather than our three [1].

The key result is that the fish discriminated colors better—its thresholds were lower, or more acute—when the target had UV contrast, meaning the UV cone was stimulated more than the other cones [1]. That is direct behavioral evidence that the UV difference is a real perceptual color for the fish, not just a physical measurement. The same logic applies to any invisible color: if a viewer with the right receptors can be trained to respond differently to it, the color is observable through behavior.

Detectors and image fusion extend the range further

Hardware can also be engineered to be sensitive in more than one band at once. A two-color infrared detector is built from stacked absorber layers separated by barrier layers, with the energy bands designed to be continuous at the interfaces so the structure can respond to two infrared bands [3]. This is a sensor-level solution: rather than filtering one broad response, the device is structured to capture distinct infrared channels.

At the image level, infrared and visible pictures can be combined. One method fuses infrared and visible images to enhance color contrast, aiming for a result that looks closer to what the eye finds comfortable during daytime viewing [4]. The fused image is a composite: visible light supplies familiar appearance, while the infrared channel adds information the eye could not otherwise receive. In each case, the invisible color becomes observable only after an instrument measures it and a display re-encodes it.

What the translated image does and doesn't tell you

A false-color or fused image is evidence about wavelength content, not a window into another animal's experience. When the anemonefish discriminates a UV target, that tells you the fish's visual system uses UV as a color channel [1]; it does not tell you what the UV "looks like" to the fish. Likewise, an infrared photothermal image showing two chemical species [2] or a pseudo-color segmentation of near-infrared differences [5] tells you where those substances or contrasts are, with the colors chosen for human legibility.

The practical answer, then, is that observing an invisible color requires either a detector tuned to that wavelength or a behaving animal that possesses the relevant receptor—and in both cases the observation arrives as a translation. The measurement is real; the color you see on the screen is a stand-in for it.

Sources used in this answer

Related research questions