SHOW SIDEBAR
What Is Near-Infrared Light? Wavelength Range, Properties and Everyday Uses

In short

• Near-infrared (NIR) light is the invisible light just beyond red, starting at about 750 nm. The upper limit is about 2500 nm in spectroscopy, and up to 3000 nm in ISO 20473.

• It behaves much like visible light (lenses and ordinary silicon camera sensors work with it), but we can't see it. That makes it ideal for sensing.

• Typical wavelengths: 850 nm and 940 nm for IR illumination, face recognition and ToF sensors; 905 nm and 1550 nm for LiDAR; 1310 nm and 1550 nm for optical fiber.

• Many black materials reflect NIR strongly, which can cause stray light in NIR cameras and sensors.

Near-infrared light is all around us. It unlocks your phone, lights up security cameras at night and helps self-driving cars see the road, but we never notice it, because our eyes can't see it. In this article we explain what near-infrared light is, its wavelength range and properties, and where it is used, with photos and video taken through our own near-infrared camera.

What is near-infrared light?

Diagram of the visible light spectrum, showing wavelengths from 380 nm (violet) to 750 nm (red)

The light our eyes can see, visible light, covers wavelengths of roughly 380 to 750 nanometers (nm). The short end (380 nm) looks violet, and the long end (750 nm) looks red. Light with shorter wavelengths than visible light is ultraviolet (UV), and light with longer wavelengths is infrared (IR).

Near-infrared is the part of infrared closest to visible light, starting right where red ends. “Near” simply means near to visible light. Because it is so close, it behaves much like visible light, yet it is invisible to us.

Near-infrared wavelength range

There is no single official boundary. Different fields divide infrared in different ways, which is why you will see different numbers in different sources.

Convention Near-infrared Other bands
ISO 20473 0.78 to 3 µm
(780 to 3000 nm)
Mid-infrared 3 to 50 µm, far-infrared 50 to 1000 µm
CIE (lighting, photobiology) IR-A: 780 to 1400 nm
IR-B: 1400 to 3000 nm
IR-C: 3 µm to 1 mm
Spectroscopy (NIR analysis) 780 to 2500 nm Mid-infrared about 2500 nm to 25 µm
Cameras and sensors About 750 to 1000 nm (NIR) SWIR about 1000 to 2500 nm, MWIR 3 to 5 µm, LWIR 8 to 14 µm

Boundaries in the camera and sensor industry vary between manufacturers. Some put the NIR/SWIR boundary at 900 nm or 1100 nm.

Infrared bands as used for cameras and sensors: near-infrared (NIR), short-wave (SWIR), medium-wave (MWIR) and long-wave infrared (LWIR), from 380 to 12000 nm
Infrared bands as named in the camera and sensor industry

The camera and sensor naming follows the detectors that can see each band. Ordinary silicon image sensors respond up to about 1000 to 1100 nm, so that range is called NIR. Beyond it (SWIR), sensors made of InGaAs (indium gallium arsenide) are needed. MWIR and LWIR are the bands used by thermal cameras.

Properties of near-infrared light

  • It is invisible. It can light up a scene for a camera without dazzling the people in it.
  • It behaves like visible light. It travels in straight lines, reflects and refracts, and ordinary lenses work with it.
  • Ordinary camera sensors can detect it. Silicon image sensors respond up to about 1000 nm, so mass-produced cameras can be used with little or no modification. Most cameras have an IR-cut filter so that photos look the way our eyes see them.
  • It is reflected light, not heat. Objects at room temperature emit infrared at much longer wavelengths (around 10 µm). NIR cameras see reflected NIR, much as an ordinary camera sees visible light.
  • It passes into some materials that block visible light. It travels a few millimeters into skin and tissue, which is used for vein authentication and pulse oximeters.
  • Things look different in NIR. Plants look bright white, and many black fabrics and black coatings reflect it strongly. Color does not tell you how a material behaves in NIR.

Common near-infrared wavelengths and where they are used

Near-infrared light sources, mostly LEDs and lasers, are made at a few standard wavelengths. These are the ones you will meet most often.

Wavelength Typical uses
700–900 nm Vein authentication, pulse oximeters and other medical sensing
850 nm IR illumination for security cameras, eye tracking, machine vision
905 nm LiDAR (detected with silicon sensors)
940 nm Face recognition, time-of-flight (ToF) distance sensors, TV remote controls
1310 / 1550 nm Optical fiber communication
1550 nm Long-range LiDAR (detected with InGaAs sensors)
780–2500 nm NIR spectroscopy: checking the sugar content of fruit, sorting plastics, analyzing food, crops and pharmaceuticals

Let's look at two of these up close with our near-infrared camera: face recognition in the iPhone, and LiDAR.

Seeing it for yourself: the iPhone's Face ID through a near-infrared camera

An iPhone, used to demonstrate near-infrared technology
A digital camera modified to capture near-infrared light

Our “near-infrared camera” is really just a regular digital camera that we modified. As explained above, its silicon sensor can pick up light up to about 1000 nm, but an IR-cut filter normally blocks everything beyond the visible range. We removed the IR-cut filter and put an IR transmission filter in its place, one that only lets through light longer than 760 nm. Ta-da! A near-infrared camera.

Now let's unlock the iPhone with Face ID. To our eyes, nothing seems to happen: the phone simply unlocks. But through the near-infrared camera…

An iPhone flashing near-infrared light during Face ID, seen through a near-infrared camera
Flash! Flash! Flash! The iPhone emits near-infrared light while Face ID is running.
The iPhone pointed at a wall, seen through a near-infrared camera
A pattern of countless near-infrared dots projected onto the wall

When we point the iPhone at a wall, it turns out that the flashing is a pattern of countless dots. The iPhone projects it while Face ID is running, or while you're making a Memoji, among other things.

The near-infrared dot pattern projected from the iPhone
The iPhone TrueDepth camera system

This is the iPhone's TrueDepth camera system. A dot projector shines a pattern of invisible near-infrared dots onto your face, and an infrared camera, mounted slightly to one side, captures it. Because the camera sees the dots from a different angle, each dot shifts sideways depending on how far away that part of your face is. By measuring these shifts, the iPhone builds a 3D map of your face. This 3D measurement method is called structured light.

Using near-infrared means the iPhone can scan your face in the dark without a visible flash. It can also work through many sunglasses, as long as the lenses let some infrared light through. Memoji uses the same system to follow the movements of your face.

Here's a video, filmed with our near-infrared camera, of TrueDepth in action.

LiDAR and autonomous driving: 905 nm or 1550 nm?

Concept image of a self-driving car dashboard

One of the key sensor technologies behind driver assistance and autonomous driving is LiDAR (Light Detection and Ranging). It works like a laser rangefinder: it fires a laser pulse and measures how long the reflected light takes to come back. By repeating this hundreds of thousands of times per second in different directions, LiDAR builds “point cloud data”, a 3D picture of the shape and distance of everything around it. Since 2020, Pro models of the iPhone have also included a small LiDAR scanner.

Image of 3D point cloud data created by LiDAR
Image of 3D point cloud data

For a car, the important questions are how finely LiDAR can sample its surroundings (so that a narrow obstacle doesn't slip between the beams), and how far it can see. At highway speeds, it needs to detect objects 100 m or more ahead. The farther away an object is, the less light comes back, so a stronger laser would help, but the laser has to stay safe for the human eye.

This is where wavelength matters. Most LiDAR uses near-infrared light at 905 nm, which inexpensive silicon sensors can detect. But light around 905 nm passes through the front of the eye and is focused onto the retina, so its power has to be kept low. At 1550 nm, most of the light is absorbed by water in the front of the eye before it reaches the retina, so eye-safety rules allow far more power, and a longer range. The catch is that silicon sensors can't detect 1550 nm, so more expensive InGaAs sensors are needed. Balancing performance against cost is one of the main questions in LiDAR design.

Why NIR matters for black materials

In near-infrared systems like these, the light that matters is light you can't see. That makes stray light easy to miss. Many materials that look black, such as black fabrics, black anodizing and ordinary black paints, reflect near-infrared light strongly. Inside a camera, sensor housing or test setup, they can cause flare, noise or false detection.

Material 550 nm
(visible)
850 nm 940 nm 1550 nm
Black anodizing (reference) 5.4% 65.7% 74.5% 87.6%
FINESHUT KIWAMI 0.70% 0.70% 0.72% 0.75%
IR Flock Sheet 0.26% 0.27% 0.30% 0.47%

Representative total hemispherical reflectance (angle of incidence 8°), measured by an external testing laboratory. Not guaranteed values. Black anodizing is not our product; we obtained a part and measured it under the same conditions for reference.

Black anodizing reflects only about 5% of visible light, but 65.7% at 850 nm. IR Flock Sheet and FINESHUT contain carbon, so they stay almost as black in the near-infrared as they are in visible light.

For photos of black materials through a near-infrared camera, and measurements for more materials, see Which Materials Reflect Infrared? What "Black" Looks Like Through a Near-Infrared Camera. For why black materials reflect NIR, see Why Black Materials Reflect Near-Infrared, and 3 Ways to Build a True NIR Absorber.

FAQ

Q. What is the wavelength range of near-infrared light?
A. Near-infrared (NIR) light starts just beyond visible red, at about 750 to 780 nm. Where it ends depends on the convention: ISO 20473 puts the upper limit at 3 µm (3000 nm), while spectroscopy usually uses 780 to 2500 nm. In cameras and sensors, “NIR” usually means about 750 to 1000 nm, the range silicon sensors can detect, and 1000 to 2500 nm is called short-wave infrared (SWIR).

Q. Can humans see near-infrared light?
A. Practically no. Our eyes are sensitive from about 380 to 750 nm, and sensitivity falls off steeply beyond red. Some people notice a faint red glow from an 850 nm LED, because part of its light is close to the visible range, but 940 nm LEDs are essentially invisible.

Q. Is near-infrared the same as thermal infrared?
A. No. Thermal cameras detect long-wave infrared (about 8 to 14 µm) emitted by objects at room temperature. Near-infrared cameras see reflected NIR light from the sun or from NIR LEDs and lasers, much like an ordinary camera sees visible light, so they cannot measure temperature at normal room temperature.

Q. Is near-infrared light harmful to the eyes?
A. Everyday sources such as remote controls, Face ID and NIR LEDs are designed to meet eye-safety standards. High-power NIR lasers, however, need care: light at roughly 400 to 1400 nm reaches the retina, and because NIR is invisible, it does not trigger a blink reflex. Light beyond about 1400 nm is mostly absorbed before it reaches the retina, which is one reason 1550 nm is used for long-range LiDAR.

Q. Does a black material absorb near-infrared light?
A. Not necessarily. Many materials that look black, such as black fabrics, black anodizing and ordinary black paints, reflect NIR light strongly. In our measurements, black anodizing reflected 65.7% at 850 nm. Materials that contain carbon, such as our IR Flock Sheet and FINESHUT, keep their reflectance low into the NIR.

First published on March 17, 2020. Revised in October 2026.

Related articles

Leave A Comment

Please note, comments must be approved before they are published