Category: SWIR

Most machine vision systems begin with visible-light cameras, but many inspection challenges remain unsolved because the information being sought simply isn’t visible in the VIS spectrum.

Many are intimidated by SWIR, worried it’s some exotic sensing technology. While it’s different in that it’s a form of vision that’s in a different part of the spectrum than VIS, we all enjoy radio and bluetooth, and we’re not natively equipped to sense those parts of the spectrum either!

Below we present SWIR as a practical problem-solving technology. And we’ve created a new area on the 1stVision website dedicated to SWIR imaging.

Sign #1: You Need to Detect Moisture or Water Content

Applications:

• Fruit and vegetable inspection
• Agricultural monitoring
• Crop stress analysis
• Food quality grading
• Moisture detection

Key takeaway: Water absorbs SWIR wavelengths, creating contrast that visible-light systems often cannot provide.

The economic payoff for fruit sorting with SWIR is to detect early bruising. That way bruised fruit isn’t shipped for display and customer rejection and poor quality – instead lightly bruised fruit can still be quickly sauced or juiced, while heavily bruised fruit may be rejected early to reduce shipping costs and preserve reputation for quality.

Sign #2: Different Materials Look the Same in Visible Light

Applications:

• Plastic sorting
• Recycling systems
• Material identification
• Foreign material detection
• Food processing

Key takeaway: Materials that appear identical in visible light often separate very clearly in SWIR.

Consider recycling and automated sorting applications: If the materials aren’t visible to humans, then neither humans nor VIS spectrum machine vision are candidates. But SWIR applications can distinguish among different types of plastic, for example, making recycling profitable or at least less costly. Also relevant for food safety applications.

---

Not sure if SWIR is right for your application?

Our machine vision specialists can review your application and recommend the right camera, lens, lighting, and filters. There is no cost for this consultation.

Sign #3: You Need to See Through Silicon

Applications:

• Semiconductor inspection
• Wafer inspection
• Die inspection
• Solar manufacturing

Key takeaway: Silicon becomes transparent at SWIR wavelengths, allowing inspection of features hidden from visible cameras.

Economics: Semiconductor and solar panel production is an expensive undertaking, and it’s essential to detect any flaws as early as possible, to reject any materials that aren’t worthy of further processing. Improving yields reduces costs and raises quality, providing competitive advantage.

Sign #4: You Need to Inspect Through Plastic Packaging

• Pharmaceutical packaging
• Medical devices
• Bottle inspection
• Seal inspection
• Fill-level verification

Key takeaway: SWIR can provide visibility through packaging materials that are difficult to inspect using conventional imaging.

Especially many plastics are opaque in VIS yet transparent in SWIR. The benefits are clear – pun initially not intended but in hindsight it works.

---

Need to inspect products through plastic packaging?

Explore our SWIR camera solutions or contact us to discuss your inspection requirements.

View SWIR cameras.

---

Sign #5: Glare, Dust, Smoke, or Haze Is Limiting Performance

Applications:

• Outdoor inspection
• Transportation systems
• Industrial process monitoring
• Harsh manufacturing environments

Key takeaway: SWIR wavelengths can often provide improved contrast and image quality in challenging environments.

How SWIR Technology Has Changed in the Last Five Years

Many engineers who looked at SWIR way back before 2020 may think of it as expensive and niche. Back then that was a fair characterization. It was still cost-effective in defined niches when the ROI made sense, but it’s fair to say it was niche.

But automatic transmission cars were once niche. As were smartphones. And streaming video. Technologies evolve.

Recent SWIR developments – two in particular:

One: Sony SenSWIR

Commercialized in just the past few years, Sony’s SenSWIR IMX992 and IMX 993 sensors continue to use InGaAs. But with innovative bonding technologies Sony has improved quantum efficiency, permitting dense pixels, smaller sensors, and the avoidance of cooling.

That’s enabled small-format cost-effective cameras such as Allied Vision’s Alvium SWIR models. Note that SenSWIR isn’t “just” SWIR (many previous SWIR cameras could only work in the SWIR portion of the spectrum), but VIS+SWIR:

• Higher resolutions
• Smaller pixels
• Faster frame rates
• More compact camera designs
• Easier machine vision integration

Two: Nextgen Quantum Dot SWIR Sensors

Allied Vision and QDI Systems recently announced a partnership to utilize Quantum Dot (QD) sensors. “The collaboration aims to accelerate the transition from traditional InGaAs-based SWIR sensors to a new class of QD sensors, delivering comparable performance while significantly improving cost, size, and weight characteristics. This breakthrough is expected to unlock new opportunities, particularly in industrial applications and drone-based imaging systems.”

What are quantum dots?

Per QDI Systems, quantum dots (QDs) are “tiny solution-processable semiconductor nanoparticles with extraordinary properties. […] At their tiny scale, quantum dots exhibit unique quantum properties, which distinguish them from larger particles. One of the most notable features of quantum dots is their size-dependent optoelectronic properties. This means that their ability to absorb and emit light can be precisely tuned by changing their size or composition.”

We’re just on the cusp of commercialized camera offerings utilizing QDs, but it’s apparent there may soon be choices beyond traditional InGaAs-only soluntion.

And that will further reduce the cost, size, and weight of SWIR cameras.

---

“Is SWIR Approaching Its CMOS Moment?”

Machine vision practitioners in the VIS space for more than a few years will recall that CCD sensors were dominant until 2007-ish and CCD still had significant market share for some years after that.

We submit that newer technologies like Sony SenSWIR and Quantum Dot SWIR sensors could, or really will, move SWIR from a niche technology into a mainstream machine vision tool

---

Choosing the Right SWIR Camera

If you are considering SWIR, we suggest you keep the following in mind. Or call us at 978-474-0044, and we’ll walk you through key choices based on your applications goals.

• Area Scan vs. Line Scan
• InGaAs vs. Sony SenSWIR vs. emerging Quantum Dot technologies
• GigE, USB3, 5GigE, CoaXPress, Camera Link, and CSI-2 interfaces
• Importance of selecting the proper camera, lens, lighting, and filters as a complete system

Or click the link for all SWIR cameras on offer at 1stVision.

---

Conclusion:

Stop asking, “Do I really need a SWIR camera?”

Start asking, “Is my visible-light camera hiding information that SWIR could reveal?”

Especially if you are into any of the following applications:

• Moisture detection
• Material discrimination
• Plastic sorting
• Recycling
• Semiconductor inspection
• Packaging verification
• Food inspection

---

Wondering if SWIR Can Solve Your Vision Challenge?

1stVision has over 30 years of machine vision experience helping customers select the right SWIR cameras, optics, filters, lighting, and software.

Send us any/all of the following. More is better but we’ll work with what you’ve got:

• Sample images
• Working distance
• Field of view
• Inspection goals

and we’ll help determine if SWIR is the right technology.

---

1st Vision’s sales engineers have over 100 years of combined experience to assist in your camera and components selection.  With a large portfolio of cameras, lenses, cables, NIC cards and industrial computers, we can provide a full vision solution!

About you: We want to hear from you!  We’ve built our brand on our know-how and like to educate the marketplace on imaging technology topics…  What would you like to hear about?… Drop a line to info@1stvision.com with what topics you’d like to know more about.

#SWIR

The JAI WAL-1001-GE and -2001- are compact GigE 1K SWIR line scan cameras with InGaAs sensors designed for high-speed industrial inspection using a high-sensitivity InGaAs sensor. While there are a number of area scan SWIR cameras on the market, line scan SWIR is a smaller niche. But if it gives you a competitive edge or innovative solution – it’s not a niche for you.

---

Optional review on SWIR and line scan

Just in case you’re coming at this from a VIS (visible spectrum) and/or area scan context…

SWIR – Short Wave Infra Red – uses a segment of the spectrum not visible to the human eye nor CMOS sensors used for VIS imaging. But SWIR is very effective at “seeing” things not revealed in VIS. For more on SWIR see our blog and/or knowledge-based article.

Line scan – unlike area scan images that capture a large 2D view and whose framerates are limited by data volumes, line scan sensors are just one or two pixels wide while hundreds or thousands of pixels long.

---

Two example applications before camera details

Here are two representative SWIR line scan applications. There are many others as well.

Semiconductor inspection: Detect flawed wafers early in production to avoid shipping flawed product, to reduce production costs, and and increase yield.

---

Plastic seal inspection: SWIR imaging reveals contrasts that are not visible with standard visible-light imaging. In this example, the heat seals in a plastic bag needed to be inspected to ensure that the seal was strong and consistent.​ The bags are produced in a continuous production process, where a single line-scan camera can capture the seams as the product moves through the system.

---

JAI WAVE WAL 1001 GE and -2001- camera highlights

For full details, including data sheets, tabular overview, etc., see both the SWIR line scan and area scan JAI WAVE cameras at our website. Here we call out just a few highlights.

1024 px at 29 kHz or 2048 px at 40 kHz:

(WAL-1001-GE and WAL-2001-GE, respectively)

Sensor layouts on 1001 vs 2001 models:

How much resolution do you need?

Key point: The 2k camera captures with both of the offset lines, this offset will reveal sub pixel details revealing smaller defects.

It will also oversample the sensor running at twice the normal speed to get the same level of detail in the movement direction. It will then output an image with the equivalent of a 2k sensor with half the pixel size. ​

Doubling down on the 2 row camera, for emphasis

While the 1k camera is already enough for many applications, let’s spend a moment longer on the 2 row WAL-2001-GE model. The biggest value proposition in this model is:

Doubled resolution for enhanced defect detection​
A 0.5-pixel offset design effectively doubles sampling density, enabling reliable detection of sub-pixel-scale defects beyond the limits of conventional 1K cameras.​

But it’s also worth noting:

Optimized cost-to-performance ratio​

The dual-row 1K sensor configuration replaces a native 2K sensor, delivering high-resolution infrared imaging while reducing overall system cost.​

…and

Internal image synthesis for simplified system design​
The 2K image is synthesized in real time inside the camera, eliminating the need for an external acquisition card and reducing both system complexity and latency.​

Resolution differences

Below, created under identical conditions, see the expected higher resolution with the 2k camera utilizing the 2 rows of pixels offset against each other.

1st Vision’s sales engineers have over 100 years of combined experience to assist in your camera and components selection.  With a large portfolio of cameras, lenses, cables, NIC cards and industrial computers, we can provide a full vision solution!

About you: We want to hear from you!  We’ve built our brand on our know-how and like to educate the marketplace on imaging technology topics…  What would you like to hear about?… Drop a line to info@1stvision.com with what topics you’d like to know more about.

#JAI

#SWIR

#Linescan

#WAL-1001-GE

#WAL-2001-GE

The Wave WAA-1300-GE-TEC combines visible and short-wave infrared (SWIR) sensitivity in a single camera, enabling enhanced material contrast, improved defect detection, and more reliable inspection results. With sensitivity from 400 nm to 1700 nm, the camera can reveal features and material characteristics that may not be visible using standard imaging technologies.

---

Applications

---

---

---

---

---

Utilize diverse spectral responses to your advantage

The application areas above are just representative, and are not meant to be exhaustive. The key point is that diverse materials provide differing spectral responses under appropriate light (including natural light).

Applications may be designed to identify and differentiate materials according to their spectral properties. Sensors, cameras, lighting, and lenses are available for machine vision applications that draw upon each of UV, VIS, IR, NIR, SWIR, MWIR, and LWIR portions of the spectrum. And combinations thereof – like the VisSWIR JAI Wave WAA-1300-GE-TEC.

---

1st Vision’s sales engineers have over 100 years of combined experience to assist in your camera and components selection.  With a large portfolio of cameras, lenses, cables, NIC cards and industrial computers, we can provide a full vision solution!

About you: We want to hear from you!  We’ve built our brand on our know-how and like to educate the marketplace on imaging technology topics…  What would you like to hear about?… Drop a line to info@1stvision.com with what topics you’d like to know more about.

#TEC

#VISSWIR

#JAI

Short Wave Infrared (SWIR) imaging is enjoying double-digit growth rates, thanks to improving technologies and performance, and innovative applications. Unlike visible-light sensors, SWIR cameras can image through silicon, plastics, and other semitransparent materials. That’s really effective for many quality control applications, materials sorting and inspection, crop management, fruit sorting, medical applications, and more.

Unlike CMOS sensors, from which high-quality images are reliably derived under wide operating conditions, SWIR sensors typically need “tuning” relative to temperature and exposure duration. First generation SWIR cameras sometimes generated images that while useful, were a bit rough and with certain limitations in the extreme. SWIR camera manufacturers have been innovating solutions to raise the performance of their cameras.

What’s the problem?

In short-wave infrared (SWIR) imaging applications, camera operation points such as exposure time, gain and bit-depth need to be adapted depending on the inspection task at hand. Image sensor defects such as defective pixels and image non-uniformities – inherent to SWIR sensors – are sensitive to the aforementioned operations points.

Unless controlled, image quality can suffer

Consider the following image:

The four parameters exposure setting, temperature, bit-depth, and gain may collectively be called the “Operating Point” of a SWIR sensor, as together they have a significant bearing on image quality. Through manual or automated adjustments, one can optimize image outcomes.

In this blog, we provide context for these concepts. And we introduce Dynamic Operating Point Optimization (DOPO) as an automated innovation available in the fx series of SWIR cameras offered by SVS Vistek / Allied Vision.

Before Dynamic Operating Point Optimization (DOPO)

SWIR cameras with some image correction capabilities – prior to DOPO we’ll describe in the next section – certainly improved image quality. Largely via defect pixel correction (DPC) and non-uniformity correction (NUC).

Defect pixel correction (DPC) is achieved by replacing the “hot” or “dead” pixel value by the average value of its nearest neighbors. As long as there isn’t a cluster defect with multiple adjacent defect pixels (typically identified and rejected at manufacturing quality control), this is an effective solution.

Non-uniformity correction (NUC) is a bit more complex, but worth understanding. The non-uniformities arise in thermal imaging due to variations in sensitivity among pixels. If uncorrected, the target image could be negatively impacted with striations, ghost images, flecks, etc.

Factory configuration of each camera, before finalizing testing and shipping, adapts for the nuanced differences among individual sensors. Correction tables are created and stored onboard the camera, so that the user receives a camera that already compensates for the variations.

---

In reality it’s a bit more complicated

In fact defect pixels aren’t always simply hot or dead: they may appear only at certain operating points (exposure duration, temperature, gain, bit-depth, or combinations thereof).

Likewise for non-uniformity characteristics.

So that factory configuration mentioned above, while satisfactory for many applications, is a one size fits all best hope compromise, relative to the tools (then) available to the camera manufacturer and the price point the market would accept. Just as with t-shirts and socks, one size doesn’t really fit every need.

Dynamic Operating Point Optimization (DOPO)

Allied Vision has introduced dynamic operating point optimization (DOPO) to further automate SWIR cameras’ capacity to adapt to changes brought about by exposure time, temperature, gain, and bit depth. Let’s examine the graphic below to understand DOPO and the added value it delivers.

First consider the Y-axis, “Image Quality”. Looking at the flat-field gray block, clearly one would prefer the artifact-free characteristics of the upper region.

Also note the X-axis, “Sensor Temperature / Exposure Time”, for an uncooled thermal sensor. (Note that some thermal cameras do have sensor cooling options, but that’s a topic for another blog.) See the black line “No correction” sloping from upper left to lower right, and how the number of image artifacts grows markedly with exposure time. Without correction the defect pixels and sensor non-uniformities are very apparent.

Now look at the gray lines labeled “NUC+DPC”. For a factory calibrated camera optimized for a sensor at 30 degrees Celsius and a 25ms exposure, the NUC and DPC corrections indeed optimize the image effectively – right at that particular operating point. And it’s “not bad” for exposure times of 20ms or 15ms to the left, or 30ms or 35ms to the right. But the corrections are less effective the further one gets away from that calibration point.

Finally let’s look at the zig-zag red lines labeled “DOPO”. Instead of the “one size best-guess” factory calibration, represented by the grey lines, a DOPO equipped camera is factory calibrated at up to 600 correction maps, varying each of exposure time, temperature, gain and bit depth across a range of steps, and building maps that represent all the stepwise permutations.

Takeaway: DOPO provides a set of correction tables not just one

So with DOPO providing a set of correction tables, the camera can automatically apply the best-fit correction for whatever operating point is in use. That’s the key point of DOPO. Unlike single-fit correction tables, with so many calibrated corrections under DOPO, the best-fit isn’t far off.

Thermal imaging with SWIR cameras – plenty of choices

There are a number of options as one selects a SWIR camera. Is your choice driven mostly by performance under extreme conditions? Size? Cost? A combination of these?

Call us at 978-474-0044. We can guide you to a best-fit solution, according to your requirements.

We might recommend a DOPO equipped camera, such as one of the fxo series SWIR cameras:

Or you might be best-served with a Goldeye camera, in cooled or uncooled models:

Or an Alvium compact camera, whether housed or modular (for embedded designs), in USB / MIPI CSI-2 or GigE interfaces.

The key message of this blog is to introduce Dynamic Operating Point Optimization – DOPO – as a set of factory calibration tables and the camera’s ability to switch amongst them. An equally important takeaway is that you may or may not need DOPO for a particular thermal imaging application. There are many SWIR options, in cameras and lenses, and we can help you choose.

1st Vision’s sales engineers have over 100 years of combined experience to assist in your camera and components selection.  With a large portfolio of cameras, lenses, cables, NIC cards and industrial computers, we can provide a full vision solution!

About you: We want to hear from you!  We’ve built our brand on our know-how and like to educate the marketplace on imaging technology topics…  What would you like to hear about?… Drop a line to info@1stvision.com with what topics you’d like to know more about.