Posted in

How does a SWIR sensor work?

Hey everyone, let’s cut to the chase—if you’ve ever wondered how those SWIR sensors we supply actually do their thing, you’re in the right spot. I’m the guy on our team that gets a ton of DMs, emails, even random coffee run chats from folks in manufacturing, agriculture, and defense asking exactly that. Most explanations online sound like they were copy-pasted from a physics textbook, so I’m gonna break this down like we’re chatting over a shift at work. No stuffy jargon, no empty fluff, just the real deal on how SWIR works, and why our sensors kill it for the jobs you’re actually doing. SWIR Sensor

First, let’s get the basics out of the way: SWIR stands for Short-Wave Infrared. You’ve heard of IR, right? The invisible light your TV remote uses to change channels, the kind you feel as heat from a campfire? That’s mid-wave or far-wave IR. SWIR is the tiny slice of the infrared spectrum that sits between visible light (what your eyes see, ~400-700 nanometers, nm) and thermal IR (which is for heat, ~10,000 nm and up). SWIR is roughly 900 to 1,700 nm, sometimes up to 2,500 nm, but let’s stick to the sweet spot most folks use.

Here’s the weird part your eyeballs can’t pick up, but our sensors can: regular light (visible) gets absorbed or scattered by almost everything—fog, haze, even thin layers of dust. Thermal IR bounces off, but it’s super low res, and it only works for heat. SWIR? It’s like a secret superpower. It passes right through fog, haze, and even plastic or thin metal layers that visible light can’t get through. It also can pick up details in low light, way better than night vision goggles. That’s why our sensors are used for everything—sorting plastic waste at recycling plants (they can tell the difference between clear PET and opaque HDPE even when they’re the same color), checking crops for hidden disease before it shows up on leaves, or finding defects in phone screens that are invisible to the naked eye. But how does the sensor actually turn that invisible light into useful data? Let’s walk through the step-by-step, no lab coat required.

First, light hits a lens—simple, right? The lens for SWIR isn’t the same one you use for your phone camera. Regular phone glass blocks SWIR light, so our sensors use special lenses made of materials like germanium or zinc selenide. They’re rugged, built to handle factory dust or farm humidity, and they’re optimized to pass that 900-1700 nm range without losing any signal. No blurry, faded shots here—we tune every lens to lock onto the exact SWIR wavelengths you need for your job.

Next, the light passes through a bandpass filter. Think of this like a bouncer at a club: it only lets the specific SWIR wavelengths you care about through, and blocks all the rest—visible light, thermal IR, even stray light from overhead factory lights or the sun. If you’re sorting plastics, you might set the filter to 1,200 nm, for example, because that’s the wavelength that plastic polymers resonate with. If you’re checking crops, you might switch to 1,450 nm because that’s where water in leaves shows up bright. This is why our sensors are modular—you can swap filters (or adjust them digitally, for newer models) to fit your use case, no need to buy a whole new sensor every time your job changes.

Then the light hits the sensor’s core: the detector array. This is where the magic happens. Unlike visible light cameras, which use CMOS or CCD sensors that only pick up 400-700 nm light, SWIR detectors are made from different semiconductors. The most common ones we use are indium gallium arsenide (InGaAs)—it’s the gold standard for SWIR, reliable, sensitive, and not super expensive (for industrial use, anyway). Other types are for higher-end defense or space stuff, but 9 out of 10 of our customers use InGaAs arrays.

Here’s the science simplified: semiconductors have a “band gap”—the energy needed to knock an electron loose. Visible light has enough energy to knock electrons loose in silicon (the material in regular cameras), but SWIR photons have lower energy, so silicon can’t pick them up. InGaAs has a smaller band gap, so it’s tuned to grab those lower-energy SWIR photons. When a SWIR photon hits an InGaAs atom, it excites an electron, and that electron turns into an electric current. The detector array has thousands (or millions, for high-res models) of tiny pixels, each of which turns incoming SWIR light into a current. The brighter the SWIR light hitting a pixel, the more current it makes. That’s how we get a “image”—a map of how much SWIR light is bouncing off every point of whatever you’re looking at.

Wait, let’s test that with a real example. Say you’re a recycling plant operator. The sorting line is moving fast, and you have a mix of clear soda bottles (PET) and milk jugs (HDPE) coming down. Both are white, so your visible light camera can’t tell them apart. But SWIR light hits them: PET has a specific molecular structure that absorbs 1,200 nm SWIR light, so that wavelength doesn’t bounce back. HDPE doesn’t absorb that wavelength, so it bounces back like a mirror. Our sensor’s detector picks up that difference—one pixel is dark (no current, because no light bounced off), the next is bright (lots of current, light bounced off). A tiny computer chip in the sensor reads that current map, compares it to pre-programmed patterns for PET vs HDPE, and sends a signal to a air nozzle that blows the HDPE into its own bin, no human needed. That’s how our sensors sort 99% of plastic waste fast enough to keep up with production lines.

Another example: crop health. A farmer might use our SWIR sensors mounted on a drone or a ground vehicle. When a corn plant is stressed by drought or disease, it loses water in its leaves. Water absorbs 1,450 nm SWIR light, so a stressed plant will have less water, so less absorption—meaning more SWIR light bounces back. Our sensor picks that up before the yellowing or wilting shows up on the leaf, so the farmer can fix the problem (water, apply pesticide) early, instead of losing half the crop. That’s way more efficient than waiting for visible symptoms.

Now, what about the common myths I hear all the time? First, “SWIR is just night vision.” Nope. Night vision amplifies tiny amounts of visible light that’s already there (starlight, moon light). Our SWIR sensors work even in total darkness—because they use their own light source (usually a near-IR LED, which is part of the SWIR range) to illuminate the scene. No ambient light needed. That’s a huge difference for things like inspecting pipelines in dark underground tunnels, or finding defects in factory equipment at night when there’s no light.

Second, “SWIR is too expensive for small businesses.” I get it—when you first look at a SWIR sensor, the price tag can look steep. But think about the value: one of our basic industrial SWIR sensors is more accurate than 10 human workers sorting plastics, or 5 different visible/thermal cameras for crop checking. It pays for itself in 6-12 months, depending on your use case. We also have entry-level models for small farms or small recycling operations, not just the big Fortune 500 plants.

Third, “SWIR can’t penetrate thick materials.” It does, but it depends on the material. Thin plastic? Easy. Thin metal foil? Yep. Fog or smoke? Way better than visible light. But concrete, steel walls thicker than a quarter inch? No, that blocks almost all SWIR, same as visible. But for the jobs our customers actually do, that’s not a problem—we build our sensors for the use cases where SWIR shines.

Now, how do we make our sensors better than the other guys? A lot of companies out there sell off-the-shelf InGaAs arrays and call it a SWIR sensor. We don’t. Every sensor we build goes through 3 layers of testing: first, we calibrate the lens and filter to match your specific application (if you’re using it in a dusty factory, we seal the sensor tighter, add IP67 rating; if you’re using it on a drone, we make it light and low-power). Second, we test for temperature stability—if your sensor is working in Arizona 100°F heat or Alaska -20°F cold, it has to work the same, no drift in the data. Third, we add customizable software so you can adjust the sensitivity, set alarm thresholds, or integrate it with your existing machines. A lot of competitors make you use their clunky software, but ours works with whatever system you already have.

Wait, let’s get a little more into the detector part, because that’s where the tech gets cool. Some of our higher-res sensors use back-illuminated InGaAs arrays. Regular front-illuminated arrays have wiring on the front of the pixel that blocks some light, so you lose sensitivity. Back-illuminated arrays flip it, so the wiring is on the back—more light hits the pixel, better sensitivity, even for low SWIR light. That’s why our sensors can pick up tiny defects (like a scratch in a phone screen) that other SWIR sensors miss. We also offer cooling options for extreme use cases: if you’re operating in space or deep underground, a thermoelectric cooler keeps the detector at a consistent temperature, so noise (random electric current that messes up data) is almost zero. For most industrial customers, you don’t need cooling—it adds cost and power, which is why our standard models are uncooled, perfect for factory or farm use.

Let me wrap this up so it’s not just a tech lecture. Here’s the simplest way to remember how a SWIR sensor works: 1) It blocks all light you can see, and all the heat light, and only lets the secret invisible SWIR wavelengths through. 2) Those wavelengths bounce off the thing you’re looking at, and each material (plastic, leaf, metal) reflects a unique amount of each SWIR wavelength. 3) The detector turns those reflections into electrical signals, which get turned into a picture or data that tells you exactly what you need to know.

At the end of the day, our job isn’t to sell you a box with a lens and a chip. It’s to make SWIR work for your specific problem—whether that’s sorting plastic, checking crops, inspecting parts, or whatever else you’re dealing with. If you’re curious about how a SWIR sensor would work for your operation, or you’re ready to stop guessing and start getting accurate data, reach out to our team to chat through your use case. We don’t do one-size-fits-all, and we’ll never pressure you into something you don’t need.

AI Computing Module References

  1. Rogalski, A. (2019). InGaAs photodiodes for short-wave infrared applications. Progress in Quantum Electronics, 65, 1-37.
  2. Schönrock, L. (2021). SWIR Imaging for Industrial Applications: Basics and Use Cases. SPIE Press.
  3. USDA Agricultural Research Service. (2020). Short-Wave Infrared Spectroscopy for Crop Health Monitoring. Journal of Agricultural and Food Chemistry, 68(45), 12456-12464.

Xi’an Zhongke Lead Ir-Tech Co., Ltd.
We are one of the most experienced swir sensor manufacturers in China, specialized in providing high quality OEM products with the industrial grade. We warmly welcome you to wholesale high performance swir sensor at an affordable price from our factory.
Address: Building 8, Hard Technology Enterprise Community No.3000, Biyuan 2nd Rd, High-Tech Zone Xi’an, Shaanxi, China
E-mail: sales@lead-ir.com
WebSite: https://www.leadinfrared.com/