Near-infrared (NIR) light covers wavelengths of the electromagnetic spectrum from approximately 780 to 2500 nanometers (nm), which is beyond human visual perception. NIR-based sensing provides machines with information about objects in the physical environment. When NIR light is emitted and reflected off of an object, an NIR sensor receives the reflected light or light pattern to gauge the distance, size, location, and identifying features of objects in the three-dimensional world.
NIR light is ideal for 3D sensing applications because it is invisible to the human eye, making it unnoticeable when cast into a user’s face, eyes, or environment. Proper filtering can remove the visible light spectrum for NIR sensing devices, increasing reliability for sensing in ambient light conditions or poor visibility conditions like rain or fog. However, because the human eye does not respond to NIR light, measuring the emissions from NIR LEDs and lasers is critical for safe and accurate 3D sensing of humans and their environments.
Radiant ProMetric® Imaging Radiometers greatly simplify the process of measuring NIR light sources by capturing complete angular light source distributions in a single image for analysis. Users of Radiant NIR measurement systems benefit from turnkey solutions combining high-resolution, CCD-based radiometers, optical components, and software test libraries for common analyses. Software tests in Radiant’s TrueTest™ Software platform include pass/fail parameters and custom output, and can be run in sequence to efficiently characterize and qualify light sources used for facial & gesture recognition, eye tracking, and other NIR-based 3D sensing applications.
Near-Infrared Light Spectrum
The NIR range covers approximately 780 to 2500 nanometers of the electromagnetic spectrum, which is outside of the visible light spectrum for human vision.
Applications by Wavelength
|- 780 nm|
- 850 nm
- 905 nm
- 930-950 nm
- 1040-1060 nm
- 1550 nm
Night Vision, Security Cameras
Short-range Automotive LiDAR (ADAS, Autonomous Vehicles)
Facial, Gesture Recognition
Terrestrial Mapping, LiDAR Systems
Long-range Automotive LiDAR (ADAS, Autonomous Vehicles)
Safety standards governing NIR light sources include:
- IEC 62471 standard for all light sources, required in America, Asia, and Europe
- IEC 60825-1 standard for lasers, including any NIR lasers
- American Conference of Governmental Industrial Hygienists (ACGIH) Threshold Limit Values (TLVs) for Physical Agents (including “lasers, light and near-infrared radiation”
Types of NIR Light Sources
- Light sources used in NIR “flood” illumination
- Light sources used in short-range proximity and time-of-flight (ToF) applications
- Light sources used to create structured light patterns (sometimes called “dot patterns”) produced by diffractive optical elements (DOE)
- Light sources used in long-range proximity sensing and ToF applications
NIR LEDs are used in devices from remote controls and facial recognition systems, to security cameras and medical instruments. LEDs are advantageous for NIR sensing because they cast a wide distribution of NIR light, covering a large field of view without gaps in 3D spatial sensing area.
Angular Distribution Measurement
Because NIR emissions can be dangerous to human vision, all angular distribution points of an NIR source must be measured. Missing any point during measurement may mean missing an irregularly strong emission that could prove hazardous to the user, especially over time.
Typically, a system called a goniometer is employed to rotate an NIR light source in front of a photodetector or a camera to capture 2D images of emissions to evaluate radiant intensity at each angle (measured in watts per steradian, W/sr). This process is time-consuming, requiring thousands of rotations to capture a complete angular measurement.
Alternatively, a radiometric camera combined with Fourier optics captures angular emission data from a single point, eliminating the need to rotate the device. The NIR Intensity Lens solution from Radiant combines a Fourier-optic lens with a 16-megapixel CCD-based radiometer to capture 3D light emissions across ±70 degrees in a single 2D polar measurement, with accuracy to 0.05 degrees per CCD pixel. This compact system reduces the cost and complexity typically required by goniometers to measure angular light distributions.
“Flood” Source Analysis
Object sensing may rely on a strong flash of NIR light (sometimes referred to as “flood” distribution) used to detect the presence the object and determine distance, even in darkness. These functions are commonly used in facial recognition to determine the distance of a face, which allows the device to accurately calculate facial curvature based on deformations of received NIR light reflections.
Time-of-flight (ToF) methods measure the distance between a device and an object based on the time difference between the emission of a signal (such as an NIR flood pulse) and when the signal’s reflection is received by the sensor. NIR light is particularly effective for ToF measurement because it can be separated from ambient light, reducing signal disturbance and improving accuracy in low light or nighttime conditions. In facial recognition applications, ToF provides highly accurate depth and spatial measurement.
Like all NIR emissions; emissions used for ToF functions must be tested to ensure they adhere to defined safety and performance parameters. Irregularities such as hot spots or a fall-off of intensity around the perimeter of the distribution need to be identified and corrected to ensure accurate proximity sensing by NIR sources.
Radiant’s NIR Lens solution provides Flood Source Analysis (distribution full width/half width), measuring the average power (intensity, maximum power, rising-edge, and falling-edge values).
Flood Source Analysis measurement for an NIR source shown in Radiant’s TT-NIRI Software module. Here, light is pulsed through a square diffuser to create a distribution with uniform intensity. The cross-section graph shows peaks of intensity and fall-off at the edges of the source’s emission pattern.
NIR emissions are also produced by vertical-cavity surface-emitting lasers (VCSELs) for applications such as optical fiber telecommunication, automotive LiDAR, and gesture or facial recognition. While NIR LEDs are typically less expensive, NIR lasers offer advantages that are driving their rapid adoption. NIR lasers are better at proximity sensing and autofocus; for example, they can be directed and reflected with greater precision for facial and hand-gesture recognition.
Because of its spatial coherence and focus, laser light can pass through small openings, making it easy to integrate and manipulate through diffractive optical elements (DOE). NIR lasers enable 3D imaging solutions with superior depth measurement and mapping capabilities, using structured light (light projected in a defined pattern) and ToF approaches for applications such as facial identification.
Structured Light Dot Pattern Analysis
To produce a structured light pattern, a single NIR laser emission is projected through an optical component called a diffractive optical element, or DOE. A DOE splits the laser into multiple emission points, which cast an array of tiny invisible dots in a grid onto a 3D object, such as a person’s face. When the light from each dot is reflected back from the object surface and received by the NIR sensor, the device calculates deformations in the pattern to determine the contours of the object. Effectively, the NIR structured light grid creates a “map” of an object’s 3D features. Structured light sensing is a common method used in biometric security such as facial recognition devices for smart phones and other protected devices or media.
The TT-NIRI Software module automatically identifies and counts each dot in a structured light pattern as a point of interest, outputting data on each dot such as inclination, azimuth, and radiant intensity (W/sr).
Measuring patterns produced by DOE ensures accurate and safe sensing of a 3D object’s unique identifiable features. Each dot in a pattern must be accurately positioned and emitted with the intended radiant intensity to ensure it is properly reflected to a device’s NIR sensor and interpreted correctly to identify the individual.
Regardless of the structured light pattern or dot position, shape, or size, Radiant’s radiometric measurement solution automatically identifies points of interest for each dot and outputs data for each dot’s angle of location (azimuth, inclination), maximum intensity, uniformity, and flux. Full dot source analytics are reported in TT-NIRI Software, including maximum peak (strongest emitter) location at x,y, maximum peak inclination/azimuth, maximum peak average, maximum peak solid angle, number of pixels at maximum peak point, spot power uniformity (between dots), total flux, and DOE flux (subtracting the background peak).