Visible light is only a tiny part of the electromagnetic spectrum. Gamma rays, X-rays, ultraviolet, infrared, microwave and radio waves each have their own unique properties and their own place in the spectrum. In this article, we’ll focus on SWIR, or short-wave infrared, which is one component of infrared (IR) light. Infrared wavelengths are those below red; the word “infra” is Latin for “below.”

SWIR Imaging Defined

As humans, we experience infrared light as mostly invisible, but we can feel it as heat. The IR spectrum is divided into different regions, and each region has different applications. The four commonly referenced regions are near infrared (NIR) with wavelengths of 750 - 1,000 nm, short-wave infrared (SWIR) at 1 - 3 μm, medium- or mid-wave infrared (MWIR) at 3 - 5 μm and long-wave infrared (LWIR) clocking in at 8 - 15 μm (Figure 1).

Figure 1. The IR spectrum is divided into different regions, and each region has different applications.

Except for the fact that you can’t see it with the naked eye, SWIR is like visible light in that photons are reflected or absorbed by an object. This is unlike mid-wave and long-wave infrared light, where heat is emitted from the object itself. SWIR imaging can highlight “defects” in an inspection that visible imaging cannot. When imaging in SWIR, water vapor and certain materials become more or less reflective, and more or less transmissive in the SWIR vs. the visible. For instance, silicon becomes transparent beyond ~1 μm, but water actually becomes more absorptive in the SWIR – especially around bands at 1.45 μm and 1.8 to 2 μm. This means colors that appear almost identical in visible light can be easily differentiated in the SWIR.

How do SWIR Vision Systems Work?

SWIR cameras are often built around Indium Gallium Arsenide (InGaAs) infrared detectors. InGaAs sensors can be made extremely sensitive, and as a result, SWIR cameras will work in light-starved conditions.

For the most part, SWIR vision systems operate in much the same way as visible systems do. You have a target, you have a light source, and a detector to capture the image. The image shows up as black and white. So, what’s the difference between imaging with monochrome cameras and SWIR imaging? Well, SWIR light is invisible to the human eye and can detect and highlight certain features that are difficult or very impossible to distinguish with visible light and visible cameras. For example:

  • SWIR can help differentiate between objects that are very similar in color in the visible spectrum;

  • SWIR can help to see through certain objects, such as silicon;

  • SWIR can help see objects with very high temperatures.

What Does SWIR Imaging Enable?

People tend to use SWIR because they can see different materials better. An example that is often used is salt versus sugar. They’re both small white crystals when viewed in the visible light, but they have quite different reflective qualities in the SWIR.

SWIR can also be used to look for water content in material, which can be beneficial for applications in agriculture, food inspection, and forestry. Any object that contains water will absorb SWIR wavelengths at one of two absorption lines – one is around 1.45 microns, another one at closer to 1.8 microns. With SWIR imaging, this enhances visibility of objects containing moisture.

With SWIR, you can generate higher contrast images in haze, mist, rain, fog, and other challenging atmospheric conditions due to less scattering as you move further out into the infrared. However, optimal effectiveness for SWIR is in a very light fog or a very light haze; with a heavy fog or haze, you can rely more on heat signatures from a thermal camera. SWIR can also detect heat, but at greater than 300 degrees Celsius. So, this is useful for looking for defects in molten glass and molten metal before it’s cooled.

Line Scan vs Area Applications

Teledyne Imaging offers both area and line scan cameras, and in June of 2020, introduced its first SWIR line scan camera. As a refresher, line scan cameras scan an object line by line as it moves on a conveyor belt or via other controlled motion, such as taking images while flying over a stationary object. This is different from area applications, or “staring” applications where you capture an image of the object.

It’s possible to make any application into a line scan application if you’re willing to move the camera or move the scene. A good example is apple inspection. You could take an image of the entire apple and process that entire image, or more effectively, you could position the apple on a conveyor belt going by a line scan camera and process the data line by line which typically results in higher resolution and processing efficiency at lower cost.

Where SWIR Shines

SWIR applications range widely from food sorting and recycling to solar panel inspection, agriculture, forestry, and surveillance. The benefits of SWIR imaging are evident. We’ll look at some of these applications and discuss how SWIR brings unique capabilities to make these tasks easier.

Food Sorting

With SWIR imaging, we can increase yields, reduce waste, and improve food quality. SWIR imaging is best used for high-value food sorting applications. As an example, SWIR would typically not be used for commodity crops like rice, but for higher value products. SWIR is more suitable for sorting tea leaves once they’re harvested and roasted. Because the tea leaves are black after roasting, it is difficult to see contaminants that may be mixed with the tea leaves. With SWIR, the quality inspection process can effectively identify stems, small stones, or other debris and eliminate them from the finished product.

Another way that we can use SWIR is in the detection of moisture content in food sorting, where moisture can show spoilage or otherwise damaged produce. For example, moisture content in fruit indicates a bruise. The bruise would be visible in SWIR before a human could detect it.

SWIR can also help to differentiate products that are similar in color (Figure 2). In this example there are cinnamon, coffee beans, rocks, and raisins. On the right you have a color image where some of those items look very similar and, on the left, you can now more clearly differentiate between these with SWIR.

Figure 2. SWIR can also help to differentiate products that are similar in color.

Recycling applications use similar sorting techniques to separate different types of materials. In plastic sorting, SWIR multispectral systems are used near the end of the sorting process because they are so sensitive. They are typically run twice or more to achieve up to 99% purity of the recycled plastic material.

Solar Panel Inspection

Since SWIR can see through silicon, another application for SWIR imagers is solar panel inspection (Figure 3). With a global push toward more sustainable sources of energy, solar panels have seen a significant increase in adoption. Manufacturers need to make sure their panels are free from defects, cracks, or saw marks on the opposite side of the wafer. In addition, SWIR can be used to identify dead spots or weak areas on a solar cell and help prove the efficacy of the cell. Overall, you wind up with a much higher quality product when SWIR is used for quality inspection. Many of these same techniques can be used in semiconductor inspection.

Figure 3. Since SWIR can see through silicon, another application for SWIR imagers is solar panel inspection.

Agriculture and Forestry

A lot of airborne imaging with SWIR is related to agricultural or forestry applications. In agriculture, farmers need to manage and respond to a seemingly endless number of challenges to ensure high quality and high yields. Weather, invasive species, and disease can wreak havoc on a crop. A farmer will see the results as a crop starts to turn yellow and wilt, but by that time it’s often too late to do anything to save the crop. With SWIR imaging, scientists can precisely monitor water absorption from the roots into the leaves and make decisions on how to treat those crops.

Data collected from imagery can also provide farmers and foresters with the insight they need to make decisions related to additional irrigation or fertilizer, helping them to manage costs.

Military Intelligence Surveillance and Reconnaissance

The military relies on SWIR for intelligence, surveillance, and reconnaissance (ISR). The US military uses SWIR for low light imaging, target recognition, and aerial reconnaissance. One way to effectively implement aerial reconnaissance is with time delay and integration (TDI); a summing technology for line scan capture where a camera is mounted under a plane and flown over an area to map it. Because photons are in short supply, the summing of multiple rows provides a clear and complete picture.

SWIR Options and the Future of SWIR

Although SWIR has many advantages, the cost of SWIR systems is still relatively high. As the technology is adopted more widely and research and development are continued, it is expected the costs will decrease.

Some companies are looking at alternatives to the InGaAs sensor for SWIR imaging. Quantum dot is one lower-cost technology, but the cost is not as low as anticipated. The biggest drawback of quantum dot technology is that it has low quantum efficiency. So, its sensitivity to light is at least a factor of four lower than InGaAs. This means in any place where photons are at a premium and the customer can afford it, InGaAs is still the way to go. From our perspective, to leverage quantum dot technology, a customer would need to be willing to give up sensitivity in exchange for lower cost. They would need to add much more illumination, which would mean additional costs. To date, there are few applications that would benefit from this tradeoff.

Sony is also releasing their first CMOS InGaAs SWIR detectors. Although based on InGaAs, they are taking InGaAs and mating it with the CMOS readout circuit, pixel by pixel, by replacing the indium metal with copper. So, it’s more of what you would call a semiconductor kind of a process, where they take in a wafer of the CMOS circuit. They put InGaAs chips on top and they actually fuse the InGaAs and silicon via these copper layers. This also allows for smaller pixels than can be achieved with indium bump bonding, which ultimately can also be about reducing cost for a given resolution.

Figure 4. Other common applications for SWIR imaging technology.

SWIR illumination is expensive – reducing pixel size can reduce sensor cost, but only if the cost of the increased illumination to be able to see something doesn’t grow faster. It’s a similar argument to why quantum dot might be cost effective but based on pixel size rather than on QE. The goal for doing that is to reduce the cost because the process is closer to a wafer level process. These are area devices, not line scan, and the pixels are very small, approximately 5 μm pixels as compared to the Teledyne DALSA Linea SWIR GigE line scan camera available as a 1k resolution camera with 12.5 μm pixels.

The SWIR technology that most people seem to feel is most promising in terms of maintaining the performance of InGaAs but reducing cost is strained layer superlattice. It’s a multi-quantum level detector and you grow different semiconductors together in different layers and engineer the band gap to give you sensitivity that corresponds to photons in the SWIR – this is possibly three to five years away.

In conclusion, SWIR imaging technology has multiple benefits, and can perform in areas where other imaging cannot. SWIR can help differentiate between objects that are very similar in color, it can help to reveal properties or defects through certain objects, and it can help to differentiate between objects at very high temperatures. Although SWIR can be costly to implement, some applications outlined here greatly benefit from its use. As future developments unfold, we anticipate even greater uses of SWIR imaging with even more cost effectiveness.

This article was written by Mike Grodzki, Product Manager, Teledyne DALSA (Waterloo, Canada). For more information, contact Mr. Grodzki at teledyne.com, or visit here .


Photonics & Imaging Technology Magazine

This article first appeared in the May, 2021 issue of Photonics & Imaging Technology Magazine.

Read more articles from this issue here.

Read more articles from the archives here.