Search
Menu
Photonics Marketplace / Cameras & Imaging / Area-Scan Cameras / Teledyne DALSA, A Teledyne Technologies Co., Machine Vision OEM Components
Full company details
Teledyne DALSA
A Teledyne Technologies Co.
Machine Vision OEM Components
Map605 McMurray Rd.
Waterloo, ON N2V 2E9
Canada
Phone: +1 519-886-6000
Fax: +1 519-886-8023

Why SWIR Imaging? Insights on Its Practical Implementations

Photonics Spectra
Jul 2021
Image sensor advancements and a groundswell of emerging applications are unveiling valuable new inspection capabilities hidden in the shortwave infrared band.

MIKE GRODZKI, TELEDYNE DALSA

Originally developed for applications in defense, infrared scanning technology targeting the shortwave infrared (SWIR) band has become more widely adopted in many other applications over the last decade. From predicting water stress or detecting disease in crops to inspecting produce headed to market, commercial SWIR imaging technology is starting to yield real benefits in the agriculture industry, including increased crop production, lower-cost produce, and reduced food waste.

Each region of the IR spectrum has different applications. Courtesy of Teledyne DALSA.


Each region of the IR spectrum has different applications. Courtesy of Teledyne DALSA.

The SWIR band is one of four commonly referenced regions, occupying the band between the near-infrared (NIR) and midwave infrared (MWIR) windows or, more specifically, the wavelengths between 1 and 3 µm. Compared to imaging sensors operating deeper into the IR, those that capture SWIR’s shorter wavelengths deliver images with higher resolution and stronger contrast, both of which are important criteria for inspection and sorting applications. SWIR imaging can also highlight features and defects that visible imaging cannot. When imaging in the SWIR range, for example, water vapor and other materials become either more or less reflective or transmissive than they would appear under visible wavelengths. Silicon becomes transparent beyond ~1 µm, while water actually becomes more absorptive in the SWIR — especially around the wavelength bands at 1.45 µm and from 1.8 to 2 µm. This allows colors that appear almost identical in visible light to be clearly differentiated in the SWIR.

Water’s strong absorption at SWIR wavelengths makes objects or features with a high moisture content appear almost black in images captured using SWIR cameras. As a result, applying an appropriate filter or light source can help to make moisture content highly evident in bruised fruit, well-irrigated crops, or bulk grains. This also enables scientists to precisely follow water absorption from a plant’s roots into its leaves or, conversely, to monitor evaporation or desiccation. These are only a handful of the inspection applications that would be difficult or impossible to perform using cameras operating in the visible range.

The SWIR region of the IR spectrum has shorter wavelengths and delivers images with higher resolution and stronger contrast. Courtesy of Teledyne DALSA.


The SWIR region of the IR spectrum has shorter wavelengths and delivers images with higher resolution and stronger contrast. Courtesy of Teledyne DALSA.

For the most part, SWIR inspection systems operate in much the same way that visible systems do: A target, a light source, and a detector capture the image. SWIR cameras are often built around indium gallium arsenide (InGaAs)-based infrared detectors, which can be extremely sensitive. As a result, SWIR cameras work well in light-starved conditions.

Farm-to-table applications

The food and agriculture industries are facing increasing demands for better quality, improved safety, and more reliable traceability. Steady hikes in production costs have also put significant pressure on growers, processors, and retailers to adapt their supply chains. As a result, the industry is implementing new developments in precision agriculture and crop modeling to increase yields and quality and to reduce costs for crop production around the world. SWIR vision systems can help to advance these goals as well.

Imaged with SWIR, objects that appear similar in the visible range can be distinguished based on their relative moisture content. Courtesy of Teledyne DALSA.


Imaged with SWIR, objects that appear similar in the visible range can be distinguished based on their relative moisture content. Courtesy of Teledyne DALSA.

The technology is gaining favor in agricultural and forestry applications due to its ability to capture details that would be difficult or impossible to detect with visible imaging systems. This is a huge benefit when it comes to monitoring nurseries, scheduling irrigation, estimating yields, and evaluating maturity, or when detecting soil salinity, crop disease, pathogens, or bruises on fresh produce. Improving farming efficiencies is certainly important, but enabling preventative and corrective action is even more critical when combating drought conditions and other threats to plant health. SWIR imaging can allow farmers to understand crop conditions and proactively address problems such as nutrient deficiency or moisture stress before they incur long-term losses. Data collected from SWIR images also provides insight into estimated crop yields, which can help to determine whether investments in extra irrigation or fertilizers are warranted and help to guide the precise application of either.

Food inspection and beyond

Once crops are grown and harvested, automation enables food processors to implement high-speed, nondestructive quality inspection and grading of produce. These capabilities contrast sharply with human-based visual inspection and grading, which are subjective, costly, and slow.

The use of SWIR imaging can further expand the benefits of automated inspection. Human inspectors, for example, often cut open and examine produce to randomly check for spoilage, a method that is wasteful. When imaged at certain bands in the NIR and SWIR, water density and distribution within the produce can be seen clearly. These images can indicate key physical attributes that help to predict other measurable qualities of produce — such as apples — including texture, water-binding capacity, and specific gravity. Abnormalities in texture and density that are not visible to the eye can indicate bruising beneath the apple’s skin. With optical sorting, bruised fruit can be effectively sorted, and waste is reduced while customer satisfaction is increased. With SWIR imaging, apples that do not meet the standard for sale in their whole form as fresh fruit can be trimmed to remove any bruised areas and sold in products such as applesauce or pie filling.

SWIR can also facilitate detection of many foreign contaminants in food products. Small sticks or stones are often difficult to visibly distinguish when they are mixed in with raisins or nuts, especially when the food being inspected is traveling at high speed down a chute. In the visible spectrum, all objects may appear brown, but when captured in the SWIR, their differences become much more obvious.

Solar panel inspection is an ideal application for SWIR imagers because SWIR can see through silicon to detect defects. Courtesy of Teledyne DALSA.


Solar panel inspection is an ideal application for SWIR imagers because SWIR can see through silicon to detect defects. Courtesy of Teledyne DALSA.

SWIR imaging is a valuable inspection tool in other industries as well. The technology’s moisture detection capabilities are key for revealing signs of leakage in packaging. The recycling industry can use SWIR imaging to distinguish and separate various types of materials. Solar panel and semiconductor inspection are other important applications for SWIR imagers, since SWIR can see through silicon. And the military continues to rely on SWIR technology for intelligence, surveillance, and reconnaissance operations.

Options for SWIR imaging

SWIR imaging’s many advantages have been offset by the cost of InGaAs cameras, which are still comparatively high versus other industrial camera technologies. However, more widespread adoption coupled with ongoing research and development will drive reductions in cost.

This pursuit of reduced costs has introduced alternatives to InGaAs sensors for SWIR imaging, such as sensors based on quantum dot technology. While nominally cheaper, the cost of quantum dot cameras is not as low as anticipated. The biggest drawback of the technology is that it has low quantum efficiency, with a sensitivity to light at least a factor of four lower than InGaAs sensors. Leveraging quantum dot-based SWIR imaging requires much more illumination to compensate for its low quantum efficiency, which would in turn require additional costs for lighting. End users need to consider such trade-offs in their application.

Another emerging sensor technology is Sony’s new CMOS SWIR detectors, which pair InGaAs technology with the CMOS readout circuit, pixel by pixel, by replacing the indium metal with copper. Borrowing from semiconductor manufacturing methods, Sony takes a wafer of CMOS circuits, applies InGaAs chips on top, and actually fuses the InGaAs and silicon via the copper layers. The goal is to reduce the cost of the conventional InGaAs architecture by leveraging a fabrication approach that is closer to cost-effective wafer processing methods.

Sony’s sensors are area-scan rather than line-scan devices, and the sensors’ pixels are very small, measuring ~5 µm. Contrast this to the pixel sizes of Teledyne DALSA’s line of Linea SWIR GigE line-scan cameras, which measure 12.5 µm in the 1K-resolution model or 25 µm in the 512-pixel model.

A SWIR technology that seems to hold the promise of maintaining the performance of InGaAs technology, albeit at a lower cost, is strained-layer superlattice sensors. These are multi-quantum-level detectors in which different semiconductors are grown together in layers and the bandgap is engineered to provide sensitivity that corresponds to photons in the SWIR. This solution, however, is potentially three to five years away.

Practical implementations

To implement SWIR imagers effectively, several important considerations and best practices need to be kept in mind. The first is lighting. These cameras work best in combination with specialized lenses and LED illumination that are compatible with SWIR imaging. To truly get the most out of SWIR, the right equipment is required.

SWIR cameras can work with conventional halogen illumination, but often this type of light requires operation at very high powers to get a usable SWIR image. High power also means high temperature, which is unacceptable in many food-related applications. LEDs are a cooler illumination source that can emit SWIR light and can effectively circumvent issues with the high power and heat, while providing a superior signal for the SWIR camera.

Second, as SWIR imaging is applied to new applications, a good deal of trial and error may be needed to optimize contrast. The availability of detailed spectral data on the material in question can be a good guide to defining wavelengths of low or high reflectance. However, predicting whether a camera will be able to detect a particular material or quality in certain conditions is difficult. Unless a specific situation has been encountered before, the best way to know is to capture sample images using a SWIR camera. Often, a hyperspectral camera is used to help single out discrete wavelengths of interest for a particular application. In addition, cooling the sensor can offer benefits. Many applications can be addressed without cooling when standard InGaAs sensors with a 1.7-µm cutoff are used. For “extended InGaAs” sensors with a cutoff wavelength beyond 1.9 µm, most applications will require active cooling. Thermoelectric cooling is often used in SWIR imaging systems to reduce dark current when integration times and temperatures are elevated. It should be noted that some customers may wish to minimize dark current to minimize random read noise, while other customers use cooling to minimize response differences caused by shifts in ambient temperature.

Finally, as with any imaging technology, trade-offs must be made when balancing dynamic range, signal-to-noise ratio, and other factors against resolution. SWIR sensors tend to be noisier and less sensitive than their CMOS counterparts. The pixels are generally larger as well, so the image may not always be of a high enough resolution for some applications. Ideally, a SWIR imager would be used in tandem with a CMOS imager, where the CMOS device provides detailed spatial information in high resolution, while the SWIR sensor identifies features that the CMOS chip cannot detect. As in visible imaging, read noise is critical for some applications. But in most applications, shot noise is still what ultimately limits customers’ ability to extract the information they seek.

The long view

The growth of SWIR imaging technology for quality inspection has delivered multiple benefits in many application areas, including agriculture and food inspection, and the list is growing. SWIR cameras can help to identify water or moisture content in crops and produce, enabling a way to see below the surface of an object to gain information that allows for more informed decision-making. SWIR imaging can also differentiate between objects that, in the visible range, are nearly identical in color to reveal properties or defects that may contaminate or adversely affect production of fresh or packaged food. Because SWIR imaging is a relatively new modality for inspection, there are many possibilities for its use in emerging markets in which visible imaging will benefit from the augmentation of what SWIR can reveal.

Meet the author

Mike Grodzki is product manager for Teledyne DALSA’s SWIR products; email: [email protected].

More news & features
We use cookies to improve user experience and analyze our website traffic as stated in our Privacy Policy. By using this website, you agree to the use of cookies unless you have disabled them.