Machine vision traditionally operates in the visible spectrum (400-700 nm) where human eyes work. Thermal imaging operates in the long-wave infrared (8-14 µm), detecting heat rather than light. Between these lies the short-wave infrared—900 to 1700 nm—a region largely invisible to both standard cameras and thermal imagers.

A SWIR camera is specifically designed to detect this band. Unlike thermal cameras, which measure emitted radiation, a SWIR camera works with reflected light—just like a visible camera. But because the wavelengths are longer, SWIR interacts with materials in fundamentally different ways.

When you deploy a SWIR camera, you are not just adding another wavelength band. You are adding a new physical sensing mechanism that sees through, differentiates, and reveals what visible light cannot.

How SWIR Imaging Works

A SWIR camera uses specialized sensors—typically InGaAs (Indium Gallium Arsenide)—that are sensitive to photons in the 900-1700 nm range. Silicon sensors (used in standard cameras) stop responding beyond about 1000 nm. InGaAs sensors are highly sensitive in SWIR, with quantum efficiency often exceeding 70%.

The imaging setup is similar to visible light: a SWIR camera requires SWIR illumination (or ambient SWIR from sunlight/tungsten lamps) and SWIR-transmissive optics. Standard glass lenses work for the near SWIR (up to ~1200 nm), but longer wavelengths require specialized lenses.

The resulting images are monochromatic (grayscale), with brighter areas reflecting more SWIR light and darker areas absorbing or transmitting it.

Critical Applications for SWIR Cameras

Semiconductor wafer inspection – Silicon is transparent at SWIR wavelengths. A SWIR camera sees through silicon wafers, revealing subsurface defects, bond line integrity, and alignment marks buried under layers. This enables inspection of bonded wafers, through-silicon vias (TSVs), and die attachment without destructive cross-sectioning.

Agricultural sorting and quality control – Water absorbs SWIR light strongly. A SWIR camera distinguishes water content in fruits, grains, and processed foods. Bruises, foreign objects (plastic, glass, stones), and moisture variations that are invisible in visible light become clearly visible in SWIR.

Pharmaceutical inspection – Different chemical compounds have unique SWIR absorption signatures. A SWIR camera can identify tablet ingredients, detect counterfeit drugs, and verify coating uniformity without laboratory analysis.

Solar cell manufacturing – Silicon solar cells are inspected using SWIR camera systems to detect micro-cracks, voids, and metallization defects. SWIR penetrates the silicon, revealing internal structure.

Plastic and material sorting – Many black plastics are opaque to visible light but transparent or semi-transparent in SWIR. A SWIR camera enables sorting of black plastics that confuse conventional optical sorters. Different plastic types (PET, HDPE, PVC) have distinct SWIR signatures.

Through-obscurant imaging – Fog, haze, smoke, and some packaging materials scatter visible light less at SWIR wavelengths. A SWIR camera can see through conditions that blind visible cameras, useful for outdoor monitoring and inspecting through translucent containers.

Medical and biological imaging – Tissue and blood absorb and scatter visible light strongly but are more transparent in SWIR. A SWIR camera enables deeper imaging for research applications.

SWIR vs. Other Infrared Technologies

Understanding how a SWIR camera compares to other technologies helps select the right tool:

TechnologyWavelengthPhysicsTypical UseVisible camera400-700 nmReflected lightGeneral inspectionSWIR camera900-1700 nmReflected lightMaterial ID, through-siliconThermal camera (MWIR)3-5 µmEmitted radiationHigh-temp monitoringThermal camera (LWIR)8-14 µmEmitted radiationPeople/equipment temperature

A SWIR camera requires illumination (like visible cameras). Thermal cameras do not. A SWIR camera sees differences in material composition and moisture. Thermal cameras see differences in temperature.

Illumination for SWIR Cameras

Unlike thermal cameras, a SWIR camera needs light. Options include:

SWIR LEDs – Available at specific wavelengths (1050 nm, 1200 nm, 1300 nm, 1450 nm, 1550 nm). Narrow band, long lifetime, instant on/off. Lower power than visible LEDs.

Broadband SWIR sources – Tungsten-halogen lamps produce strong SWIR output but generate heat and have shorter lifetimes.

Laser illumination – For long-range or structured light applications, SWIR lasers provide high intensity.

The choice depends on the application. For water detection, 1450 nm is optimal (strong water absorption). For through-silicon inspection, 1200-1300 nm provides good penetration.

Optics for SWIR Cameras

A SWIR camera requires compatible optics:

Lens materials – Standard glass works to about 1200 nm. For full SWIR range (up to 1700 nm), lenses must use special glasses, quartz, or CaF₂. Many C-mount SWIR lenses are available.

Filters – Bandpass filters isolate specific SWIR wavelengths. Long-pass filters block visible light to prevent blooming. Cold mirrors reflect SWIR while transmitting visible.

Window materials – Standard glass absorbs SWIR. System windows must be SWIR-transmissive (sapphire, quartz, specific glasses).

Sensor Cooling and Sensitivity

SWIR imaging has unique noise considerations:

Dark current – InGaAs sensors generate dark current that increases with temperature. For high-sensitivity or long-exposure SWIR camera applications, cooling is essential.

Uncooled SWIR cameras – Work well for moderate frame rates and bright illumination. Lower cost, larger, consume less power.

Cooled SWIR cameras – Use TEC (thermoelectric) or even cryogenic cooling for lowest noise. Required for weak signals or very long exposures. Higher cost, larger, consume more power.

Most industrial SWIR camera applications use uncooled sensors with active temperature stabilization.

Resolution and Frame Rate Considerations

SWIR camera technology has historically lagged behind visible sensors in resolution and speed. Modern sensors are catching up:

• VGA (640 x 512) is common

• 1.3 MP (1280 x 1024) available at higher cost

• 5 MP sensors emerging

Frame rates from 30 fps to 300+ fps depending on resolution and interface. For high-speed SWIR inspection, expect higher costs.

Industries Leading SWIR Adoption

Semiconductor manufacturing – Wafer bonding, TSV inspection, die attach verification. SWIR is becoming standard in advanced packaging lines.

Recycling and material sorting – Black plastic sorting, polymer identification, metal/glass removal. A SWIR camera adds material discrimination capability that visible systems lack.

Food and agriculture – Moisture detection, foreign object removal, ripeness sorting. SWIR provides nondestructive internal quality assessment.

Pharmaceuticals – Counterfeit detection, coating verification, ingredient identification. Regulatory pressure drives SWIR adoption.

Security and surveillance – Through-fog imaging, camouflage detection, long-range identification. A SWIR camera sees what visible and thermal miss.

Selecting Your SWIR Camera

When choosing a SWIR camera, specify:

• Spectral range (standard 900-1700 nm or extended to 2200 nm)

• Cooling requirement (uncooled or TEC-cooled)

• Resolution (VGA, 1.3 MP, or higher)

• Frame rate (fps needed for your line speed)

• Interface (GigE, USB3, Camera Link)

• Lens mount (C-mount or custom)

Also verify illumination compatibility. A SWIR camera is only as good as its light source.

The Future of SWIR Imaging

SWIR camera costs have dropped significantly, making the technology practical for mainstream industrial inspection. As sensor fabrication scales and InGaAs-on-silicon becomes more common, SWIR will become as routine as visible imaging for many applications.

For inspections that require seeing through silicon, identifying materials, or detecting moisture, a SWIR camera is not an exotic specialty tool—it is the right tool. And as costs continue to fall, the question will shift from "Can we justify SWIR?" to "Can we afford not to use it?"

Privacy Terms Contact Us About

© 2026 WriteXO. A