SWIR in Industrial Machine Vision: Fundamentals, Challenges, and Applications

Today, industrial machine vision utilizes far more than just the light visible to the human eye. Alongside standard visible-spectrum camera systems, SWIR (Short Wave Infrared) systems are rapidly gaining traction. They enable applications that are impossible or highly restricted with conventional cameras—such as moisture detection, package content inspection, or the analysis of specific materials.
 

SWIR is often referred to as "invisible visible light." The reason is simple: unlike thermal imaging cameras, SWIR cameras primarily capture reflected radiation. Therefore, they generate images using a principle very similar to visible-spectrum cameras. Although the human eye cannot perceive these wavelengths, SWIR can reveal information that remains completely hidden in visible light.

It is exactly this combination of classic imaging and supplementary material data that makes SWIR so appealing for industrial applications. To fully grasp the advantages of SWIR systems, it helps to first look at where this spectral range fits in.

^{CUPRITE SWIR lens from Schneider-Kreuznach}

While the exact division of the electromagnetic spectrum isn't officially standardized, a general consensus has been established in science and technology.

General classification of the electromagnetic spectrum:

• Visible Light (VIS): 380–780 nm
• Near Infrared (NIR): 780–1400 nm
• Short Wave Infrared (SWIR): 1400–3000 nm
• Mid Wave Infrared (MWIR): 3–8 µm
• Long Wave Infrared (LWIR): 8–15 µm
• Far Infrared (FIR): 15 µm–1 mm

The visible range covers the light humans can see. Near Infrared follows immediately after. The boundary between NIR and SWIR is often defined by the strong water absorption line at approximately 1450 nm.

However, in industrial machine vision, the categorization is more closely aligned with the properties of common sensors. These ranges are determined by the specific light-sensitive materials used in the sensors.

Sensor-based classification for industrial machine vision:

• Visible Light (VIS): 380–780 nm
• Near Infrared (NIR): 780–900 nm
• Short Wave Infrared (SWIR): 900–1700 nm
• Extended SWIR (eSWIR): 1700–2500 nm
• Mid Wave Infrared (MWIR): 3–8 µm
• Long Wave Infrared (LWIR): 8–15 µm
• Far Infrared (FIR): 15 µm–1 mm

This is the standard classification used to market cameras and lenses today.

Selection of Schneider-Kreuznach SWIR lenses for different sensor formats

A key characteristic of both NIR and SWIR is that the radiation is predominantly reflected by objects. Therefore, just like in the visible range, cameras capture reflected light. However, as the wavelength increases, an object's own thermal emission becomes more prominent. From about 3 µm onwards, thermal radiation becomes relevant for many applications, eventually dominating the MWIR and particularly the LWIR bands. So, while SWIR cameras mainly process reflected light, thermal infrared cameras primarily detect the heat emitted by the objects themselves.


Why is SWIR So Crucial for Industry?

Under visible light, many materials look identical and are incredibly difficult to differentiate with a standard camera. But in the SWIR range, the optical properties of many substances change dramatically. Different materials reflect or absorb SWIR radiation to varying degrees, creating strong contrasts that simply don't exist in the visible spectrum.

This exact trait makes SWIR invaluable for industrial machine vision. Materials that look indistinguishable to the human eye or a standard camera can be clearly separated. Simultaneously, hidden properties like moisture content, material composition, or internal structures are brought to light.

SWIR cameras operate in a spectral range where many materials exhibit entirely different optical behavior compared to visible light. Some opaque materials become transparent, while others become highly absorptive. The behavior of water is particularly fascinating: while it appears mostly transparent in visible light, it absorbs SWIR radiation highly effectively. As a result, water-rich areas appear starkly dark. This effect is widely leveraged in food inspection and agricultural applications.

Possible uses for accessories when using a CUPRITE SWIR lens.

SWIR Sensors: Higher Resolutions Through Smaller Pixels

Various sensor technologies are available today for image acquisition. Depending on the model, common SWIR sensors cover sensitivity ranges between approximately 950 nm and 2550 nm. In addition, there are sensors that capture both the visible (VIS) and SWIR spectra and operate, for example, from 400 nm to 1,700 nm.

Significant progress has been made in sensor development in recent years. While early SWIR sensors had pixel sizes of about 20 µm, modern sensors now measure around 3.5 µm. With smaller pixels, the achievable resolutions increase significantly, making SWIR systems increasingly attractive for traditional industrial image processing tasks.

It is expected that future sensor generations will enable even smaller pixel sizes while covering broader spectral ranges. However, this also places greater demands on the lenses used.

Another important aspect is the spectral optimization of the entire system. The lens’s transmission should match the sensor’s sensitivity as closely as possible. If necessary, additional filters are used (e.g., UV/IR-cut) to suppress unwanted spectral components—such as residual light from the visible spectrum. Because regardless of sensor quality, the same principle applies in the SWIR range: without sufficient light, no usable image can be produced.

^{SWIR filters from Schneider-Kreuznach}

While standard lenses designed for visible light often allow NIR or SWIR light to pass through, this by no means guarantees they will deliver good optical performance outside the VIS range.

Avoiding Focus Shift and Optical Aberrations

The primary reason lies in the wavelength-dependent refraction of light. Every wavelength has its own specific focal point. This phenomenon is known as "focus shift."

If a lens originally designed for the visible spectrum is used in the SWIR range, the focal plane usually shifts significantly. The resulting image can become blurry, and overall image quality noticeably degrades.

However, the most severe limitations relate to the achievable optical performance. A lens only reaches its maximum imaging quality within the wavelength range it was specifically engineered for. Using a VIS lens in the SWIR range will result in a drastic drop in performance and introduce significant optical aberrations, such as spherical aberration, astigmatism, or coma.

Designing lenses that simultaneously cover a massive spectral range—for instance, from 400 nm to 1700 nm or beyond—is exceptionally challenging. The wider the spectral range, the more complex the optical design must be to ensure the focal points of all wavelengths remain as close together as possible.

For this reason, modern SWIR lenses utilize specially selected glass types and coatings optimized for the target wavelengths. Furthermore, the optical design itself is custom-tailored to the SWIR range. This minimizes aberrations and yields vastly superior image quality compared to a classic VIS lens operating outside its intended scope.

As high-resolution SWIR sensors continue to evolve, this aspect will become even more critical. Smaller pixel sizes demand higher resolution and absolute precision from lenses. Experienced precision optics manufacturers address these demands with their customary diligence, translating them into market-ready lenses.

CHAROITE

More

CUPRITE

More

CRYOLITE

More

^{Diagram illustrating chromatic aberration using light rays refracted by a lens.}