Short Wave Infrared Imaging Technology and Principles

The shortwave infrared (SWIR) range occupies the electromagnetic spectrum just above the near-infrared. Between 1,050 and 2,500 nm – the shortwave infrared range lies far beyond the detection capacity of standard silicon-based imaging sensors.

Despite this drawback, the unique inspection, sorting, and quality control capabilities of the SWIR band has seen it increasingly used in machine vision, in addition to ambient light applications such as surveillance.

The quantum efficiency of silicon rapidly fades beyond 800 nm, so because of this SWIR sensors tend to build on alternative chemical compositions, including indium gallium arsenide (InGaAs) and mercury cadmium telluride (MCT). More modern SWIR imagers are also leveraging individual sensor architectures, like quantum dot technology. The right image modality depends on its application, which makes selection vital.

Each of these technologies dramatically increases imaging capabilities beyond visible sensors – both through the extended spectrum they open up as well as the unique ways that SWIR light alters the presentation of familiar materials.

The mid-wave IR (MWIR) spectrum is where MCT cameras are more effective.  SWIR wavelengths that lie nearer this spectrum take on the ability of that spectral region to capture energy which is emitted from an object. The photons in this range are less susceptible to Rayleigh scattering caused by particles of smaller diameter, as they are comparatively longer in wavelength. This means that SWIR imagers can see through smoke, haze, or fog.

SWIR wavelengths which are shorter - between approximately 900 nm to 1,700 nm — act similarly to photons which are in the visible range. Despite the fact that the spectral content of targets in SWIR differs, the images produced remain more visual in their characteristics and less akin to the lower resolution thermal behavior of the MWIR and LWIR light bands.

This quality renders them more closely aligned with the needs of many industrial machine vision applications. SWIR’s shorter wavelength, as compared to MWIR and LWIR, facilitates higher-resolution images with stronger contrast; both significant criteria for sorting and inspection.

SWIR occupies a spot in the non-visible light spectrum between near infrared (NIR) and longwave IR. It behaves more like visible light than the thermal energy of the IR spectrum.

Figure 1 - SWIR occupies a spot in the non-visible light spectrum between near infrared (NIR) and longwave IR. It behaves more like visible light than the thermal energy of the IR spectrum.

Furthermore, although the cameras operating at the shallow end of SWIR possess similar light capture techniques to visible cameras, the images between these and those created with silicon sensors are very different, even when imaging the same item.

Both physics and chemistry are behind this phenomenon. Interaction of light and matter involved an energy transaction. When electromagnetic energy transfers to an object’s molecules, it is absorbed by the object’s surface absorbs. If this does not happen, the energy is reflected. Materials that appear similar at one wavelength will appear entirely different at another, as each discrete wavelength has its own unique energy definition.

Thus, the unique capability of SWIR cameras can capture these high resolution images of familiar items. These images will appear completely different to a conventional silicon imager operating in the visible range.

SWIR Detectors Arrayed

In the 900 to 1,700 nm window of the SWIR range, InGaAs sensors are currently the prevailing camera technology. Compared to other SWIR imaging modalities, they are comparatively cost-efficient and mature, rendering them the most commonly used technology in machine vision applications involving sorting, inspection, and quality control.

InGaAs sensors provide high detection performance and fast response speeds, much like silicon-based detectors operating in the visible range, despite the fact that their photo sensitivity is dependent on wavelength. The fact that they are solid-state devices which incorporate no shutters or other moving parts makes them resistant to vibration, which is common on factory floors. InGaAs devices do not need expensive silicon or germanium lenses to leverage conventional glass optics, in contrast to SWIR cameras targeting thermal imaging applications.

Generally, InGaAs cameras which target SWIR applications in industrial machine vision do not need cooling. Nonetheless, cooling the sensor can significantly reduce dark current for improved image quality and, in some applications, longer exposure time, as demonstrated by this comparison of three Hamamatsu InGaAs cameras.

Typical InGaAs SWIR Camera Performance.

Figure 2 - Typical InGaAs SWIR Camera Performance.

InGaAs versus Quantum Dot Structure.

Figure 3 - InGaAs versus Quantum Dot Structure.

It is key to note that cameras built on quantum dot technology are also gaining traction as a comparatively new SWIR imaging technology. These devices operate in a spectral band which overlaps that of InGaAs sensors, which makes quantum dot cameras a direct competitor to the existing technology.

The lower quantum efficiency (QE) of quantum dot-based cameras is one aspect to consider in comparison to InGaAs imagers. As this leads to lower camera sensitivity, this could be considered a drawback. However, this sensitivity may not be as limiting as first thought, with SWIR lighting in controlled machine vision applications as well as the fact the QE can also reasonably be expected to improve as the technology matures.

Due to the relative novelty of quantum dot cameras, they tend to have higher costs attached; however, this too can be expected to decrease with the maturation of the technology. The same can be said for InGaAs cameras, too: as interest in SWIR’s potential for machine vision grows, improved economies of scale, manufacturing techniques and better yields will all be contributing factors that lessen the cost of both camera technologies.

Different Is Better

SWIR’s longer wavelengths as compared to those in the visible range see very different interactions with atomic structures, which offers some new and unique imaging possibilities for use within machine vision applications. When imaged in the SWIR spectrum, familiar items appear very differently, so emphasizing the industrial machine vision aspect has enabled myriad applications which would be difficult or impossible with visible lighting and cameras.

While the bandgap of silicon molecules makes the material absorb visible and NIR wavelengths, for example, silicon transmits lower energy SWIR wavelengths which causes semiconductor wafers to become transparent in this spectral range. This broadens options for raw material industrial application, like imaging defects both inside of and on the surface of silicon wafers. This quality of SWIR light also allows fiducial marks for alignment to be seen through the back surfaces of both wafers, improving accuracy, which also benefits wafer bonding applications.

The inspection and sorting of product is one of SWIR’s most promising machine vision applications. Water appears almost black in images of objects illuminated at that wavelength, as it is highly absorbent at both 1,450 nm and 1,900 nm. Due to this, applying an appropriate light source or filter can help display clearly the moisture content in bruised fruit, bulk grains, or well-irrigated crops.

Silicon turns translucent and passes SWIR light. This property aids many machine vision applications related to the semiconductor manufacturing process.

Figure 4 - Silicon turns translucent and passes SWIR light. This property aids many machine vision applications related to the semiconductor manufacturing process.

When fruit is bruised, the cell walls break down and the area develops a higher moisture content. Water absorbs many bands of light in the SWIR range. This absorption allows SWIR imaging to see bruises that aren’t detectable by visible camera technologies.

Figure 5 - When fruit is bruised, the cell walls break down and the area develops a higher moisture content. Water absorbs many bands of light in the SWIR range. This absorption allows SWIR imaging to see bruises that aren’t detectable by visible camera technologies.

A 1550 nm SWIR light can enable a SWIR camera to see through a plastic continer and show the fluid level.

Figure 6 - A 1550 nm SWIR light can enable a SWIR camera to see through a plastic continer and show the fluid level.

The value of moisture detection expands much further than bruised produce, however. This SWIR imaging can analyze whether dyed textiles or particle board are dry enough for further processing. This imaging can also examine the seal integrity and quality of packaging of goods, particularly if high-moisture items are contained within.

Manifold plastics that appear opaque at visible wavelengths become translucent in the SWIR range. This translucency provides new methods for inspecting product volume inside sealed plastic containers. The ability of SWIR light to penetrate plastic also offers multiple ways in which the fill levels of pharmaceuticals dispensed in white plastic bottles can be inspected.

Whilst the term “plastic” may be applied to multiple polymer chemistries that appear alike in the visible range, SWIR light illuminates the key differences between materials and permits easy identification. This quality is useful in recycling applications that leverage SWIR cameras operating from 1,100 to 2,200 nm, as they can be used to identify different polymers on a sortation conveyer.

The unique interaction of SWIR light’s with manifold materials is currently at the beginning of its potential.  Though it is not predictable how SWIR wavelengths will illuminate and image complex chemical compositions, such as pharmaceuticals, it is clear that the potential implications are vast and wide-ranging.

Illumination Beyond the Visible

Inspection and other operations often benefit from active illumination within narrow bands to promote higher contrast of objects and features, as with machine vision applications in the visible range.

LEDs contained within the SWIR range have typically been low output and spanned relatively broad ranges in the spectrum, to this day. Recent advancements in technology have permitted higher outputs in more controlled, narrow spectrums. SWIR LEDs have reached a stage at which they are controlled and bright enough for imaging, though they will remain lower output and broader spectrum than their equivalents in the visible spectrum.

These light sources, which emit at peak wavelengths of 1,050, 1,200, 1,300, 1,450, and 1,550 nm, now provide high enough powers to offer potential new options for machine vision lighting.

LED sources which operate in the SWIR range are comparable in configuration to those used in conventional visible-range machine vision applications, with this new technology, similar in comparison to familiar visible LED lighting.  These sources may be strobed individually or combined, much like visible-light LEDs, and used to enable similarly complicated image capture.

SWIR LED lighting is available in many common form factors and easy to use. There is a good selection of wavelengths from 1050-1550 nm.

Figure 7 - SWIR LED lighting is available in many common form factors and easy to use. There is a good selection of wavelengths from 1050-1550 nm.

Conclusion

The increased yields of InGaAs sensors and developments of new SWIR imaging technologies will keep the technology costs down, and increase the accessibility of these options for machine vision integrators. The greatest barrier to wider adoption of SWIR imaging may simply be the lack of research.

Though broadband light sources are useful for some applications in this range, capturing data within specific bands often provides more applicable or suitable image data. The key question to be asked is what wavelength will be most effective for their unique application – and of course, the SWIR spectrum possesses many more wavelengths than just the visible. Even advanced tools, like hyperspectral spectrometers must go through trial and error to answer this question.

However, there remain many positives. Broader imaging options lie in store when the user has determined the optimal SWIR wavelength for their own application. It is no longer necessary to rely on filters or algorithms to maximize poor imaging schemes when narrowing the spectrum.

SWIR LEDs will be able to deliver intense illumination at a wide selection of SWIR wavelengths, and thus integrators can have confidence that they will be able to match the optimal light source to their camera and application. Ultimately, many will be able to capitalize on the possibilities that await them in the SWIR spectrum.

This information has been sourced, reviewed and adapted from materials provided by Smart Vision Lights.

For more information on this source, please visit Smart Vision Lights.

Citations

Please use one of the following formats to cite this article in your essay, paper or report:

  • APA

    Smart Vision Lights. (2019, July 24). Short Wave Infrared Imaging Technology and Principles. AZoSensors. Retrieved on December 11, 2019 from https://www.azosensors.com/article.aspx?ArticleID=1694.

  • MLA

    Smart Vision Lights. "Short Wave Infrared Imaging Technology and Principles". AZoSensors. 11 December 2019. <https://www.azosensors.com/article.aspx?ArticleID=1694>.

  • Chicago

    Smart Vision Lights. "Short Wave Infrared Imaging Technology and Principles". AZoSensors. https://www.azosensors.com/article.aspx?ArticleID=1694. (accessed December 11, 2019).

  • Harvard

    Smart Vision Lights. 2019. Short Wave Infrared Imaging Technology and Principles. AZoSensors, viewed 11 December 2019, https://www.azosensors.com/article.aspx?ArticleID=1694.

Comments

  1. zafrina zara zafrina zara India says:

    it is amazing.  We also inplant training in chennai for ece students in the domain embedded system using sensors.

The opinions expressed here are the views of the writer and do not necessarily reflect the views and opinions of AZoSensors.com.

Tell Us What You Think

Do you have a review, update or anything you would like to add to this article?

Leave your feedback
Submit