Category: Cameras

Posted on

5 Signs Your Machine Vision Application Needs SWIR Imaging

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

Drone mounted SWIR sensor identifies moisture content in crops – could be used to determine irrigation needs – Courtesy Tamron

Applications:

  • Fruit and vegetable inspection
  • Agricultural monitoring
  • Crop stress analysis
  • Food quality grading
  • Moisture detection
VIS image left; SWIR image right reveals moisture content – Courtesy Allied Vision

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.

VIS left; SWIR right: differentiate materials with SWIR – Courtesy Allied Vision

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.

Sign #3: You Need to See Through Silicon

VIS image (left) only finds the large crack; SWIR image (right) also finds smaller upper crack – Courtesy JAI

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

SWIR clearly reveals seal quality better than VIS – Courtesy JAI
Left: VIS image of same scene Right in SWIR – Courtesy QDI Systems
  • 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.

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

Note how SWIR “sees through” smoke – Courtesy Sony

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:

Courtesy Sony
  • 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.

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

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

Posted on

How to be successful in line scan applications

“A line scan system is only as good as its weakest component”

Granted, any system is only as good as its weakest component. But with some systems one can cut corners here or there and get away with it. With line scan mostly not. But the rewards of getting it right can deliver effective applications with excellent return on investment.

Some key line scan systems components – Photo credits clockwise from top left: Kowa, JAI, Shutterstock, Teledyne DALSA.

Success depends upon balancing all of the following:

  • Camera 
  • Lens 
  • Lighting 
  • Motion system 
  • Encoder 
  • Mechanical stability 
  • Image processing 

Below we’ll look at each of these in turn.

Camera (and the sensor it contains)

One selects a line scan camera containing typically “one narrow row” of pixels. The row’s length is chosen to correspond to the width of the thing you are inspecting. The other dimension for the 2D application comes via the motion of the object past the camera – or vice versa.

Left: A representative line scan camera – Courtesy Teledyne DALSA

The “narrow row” may be 1, 2, or sometimes 3 or 4 pixels deep. A single pixel row offers the lowest cost, and can be sufficient for some applications. Adding a second or third pixel row, typically offset or staggered with respect to the adjacent row, offers improved resolution.

Line scan cameras may be monochrome or color in the visible (VIS) range. Or SWIR in the infrared portion of the spectrum. Or blended VIS-SWIR.

Right: Representative line scan camera front and back sides – Courtesy JAI.

There are several tens of line scan cameras from which to choose.

Lens

The lens focuses the target object onto your camera’s sensor, so the photons are captured in the pixel well and digitized for subsequent interpretation. Don’t cut corners on the lens.

Certain lenses are optimized for line scan applications. You may or may not need that. Call us at 978-474-0044 so we can help you make a good choice.

Left: LF high-resolution large-format machine vision lenses – Courtesy KOWA.

Lighting

Successful image processing depends on sufficient contrast to differentiate amongst details. While the sensor and the lens are critical, so too is lighting. In fact good lighting can sometimes permit using a “lesser” sensor or lens. Or achieving ambitious outcomes otherwise not possible without attention to lighting.

See our Knowledge Base Techbriefs on lighting. Each may be downloaded for free.

Motion system

As noted above in the camera segment, the other dimension for the 2D application comes via the motion of the object past the camera – or vice versa. This is the key aspect of line scan systems – there’s a never ending dimension in the direction of motion, and each image is a time-slice from that continuous movement.

One can move the “product” along a conveyer belt, “web” of spooled paper, etc., while holding the line scan camera and lighting fixed. Or one can move the line scan camera across the product, as with a flatbed scanner.

Ideally the motion should be continuous speed – or if variable use an encoder (see below). Capable of long use without a lot of frequent re-calibration beyond scheduled service maintenance. And with little variability in the third dimension – unless of course your application is a blend of line scan and height profiling, such as laser profiling.

Encoder

For variable speed product motion, e.g. conveyor belt with non-uniform speed during imaging, an encoder converts the rotary motion of a gear or transport wheel into electrical signals correlated to the speed of rotation.

Those signals can drive the shutter trigger on the camera.

Image royalty free – Courtesy Shutterstock.

Bonus blurb – calculating line rate

Related to how to make the whole system come together… and certainly tied to the motion system and encoder segments above, see our blog from some months ago on how to calculate line rate based on conveyor speed. That blog includes a downloadable spreadsheet with labeled fields and examples… and embedded formulas.

Mechanical stability

There’s no obvious graphic to show for this one, but one doesn’t have to be a mechanical or an optical engineer to get the point. Unplanned motion is highly undesirable, as all those lens and sensor choices are based upon a stable environment.

Sure, motion in one dimension is a given with line scan, but fast shutter takes care of that. But “wiggle” in the dimension of the pixel line, or height variability of the product being inspected – or of the line scan camera itself, introduce blur and uncertainty. Or the need to compensate for it.

Image processing

Machine vision is all about letting software interpret the image, adding value through speed, accuracy, tirelessness, and consistency. From camera configuration and control through image processing, you have a range of options.

There are camera manufacturer SDKs, third party software, and open source choices. And OS options including Windows, Linux, and various embedded systems platforms.

Sapera LT SDK windows – Courtesy Teledyne DALSA

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.

#linescan

Posted on

JAI WAL-1001-GE and 2001 SWIR line scan cameras

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.

JAI WAL-1001-GE SWIR line scan camera – Courtesy JAI
SWIR sees things that VIS cannot – Courtesy JAI

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.

VIS image (left) only finds the large crack; SWIR image (right) also finds smaller upper crack – Courtesy JAI

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.

SWIR clearly reveals seal quality better than VIS – Courtesy JAI

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)

1k at 29 kHz or 2k at 40 kHz – Courtesy JAI

Sensor layouts on 1001 vs 2001 models:

Single row of pixels on 1001, vs. 2 stagger-offset rows on 2001 – Courtesy JAI

How much resolution do you need?

Resolution outcomes from 1001 vs 2001 models – Courtesy JAI

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.

Comparing resolution outcomes on an intentionally challenging example – Courtesy JAI

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

Posted on

Camera interfaces update

Machine Vision Interfaces

In 2025 we released a blog and a revised TechBrief on machine vision camera interfaces. Things continue to evolve, as 25 GigE and 100 GigE cameras are now available. So here we call out those new entries to our updated TechBrief Camera Interfaces Explained, available for free download from our Knowledge Base.

Machine Vision Interfaces
Some common machine vision interfaces

Camera interfaces often come to the market as derivative or augmented standards from computer networking. This is certainly the case as ethernet evolved to GigE, 5 GigE, 10 GigE, and now 25 and 100 GigE, respectively. The networking pioneers tend to be the early adopters, and the cable manufacturers and interface designers develop the gear while the software and firmware leaders refine the protocols.

The global machine vision industry, while valued annually at more than $20 billion USD, benefits by riding the networking coattails, typically a few years behind the networking early adopters. That has the advantage that the protocols have matured and the third-party cables are on the market. So the machine vision standards developers – and the camera adopters inherit all that’s come before.

Seeing is believing: Note resolution and FPS columns below

In the 25 GigE and 100 GigE collections below, note both:

  1. Some large sensors at impressive frame rates – e.g. 105 MP at 112 fps
  2. Some medium sized sensors at incredible frame rates – e.g. 12 MP at 660 fps
25 GigE cameras in 1stVision’s portfolio
100 GigE cameras available at 1stVision

An interface for every application requirement

Whether your application needs only a few frames per second, from a modest sized sensor, with the camera near the PC host – or bandwidth intensive requirements – possibly at significant cable length – there’s likely an interface suited to your needs. Download the updated free TechBrief.

Alternatively, start with camera selection, and let that drive the interface. That method works too, as camera manufacturers typically design-in the interface optimal to move the data volumes off the sensor, for the type of applications they designed the camera for. But sometimes the same sensor is found with two or more interface options – so it’s helpful to be “interface-aware” when selecting among cameras.

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.

#25GigE

#100GigE