SWIR Camera Lens for Industrial Moisture Detection: How 900–1700 nm Imaging Reveals Water-Content Differences

Industrial moisture detection is fundamentally different from ordinary visual inspection. A visible camera can identify colour, shape, surface damage or contamination that produces obvious visible contrast, but it cannot reliably determine whether two visually similar regions contain different amounts of water. SWIR imaging changes the inspection mechanism from appearance-based detection to wavelength-dependent material response. Within the 900–1700 nm region, water exhibits characteristic absorption behaviour, with particularly strong absorption around 1450 nm. As water content increases, less incident radiation around this absorption region may return to the camera in a reflectance configuration, creating measurable intensity differences between wetter and drier regions. The approximately 1450 nm water-absorption feature lies directly inside the operating range of the Kyptec Automation® SWIR Camera Lens portfolio.

That principle makes an industrial SWIR camera lens for moisture detection relevant to food inspection, agricultural sorting, drying-process verification, pharmaceutical processing, textile inspection, paper or web inspection, packaged-product analysis and many other processes where water content influences quality but cannot be judged reliably from visible appearance alone. The lens, however, does not measure moisture by itself. Its job is to transmit and form a sufficiently sharp, stable image from the useful SWIR wavelengths onto the camera sensor. Reliable moisture classification then depends on the complete chain: wavelength selection, illumination stability, optical transmission, sensor response, exposure, field of view, surface geometry, calibration and image-processing logic.

The dedicated Kyptec Automation® SWIR Camera Lens range covers 900–1700 nm and currently includes 8.5 mm, 12.5 mm, 25 mm, 35 mm and 50 mm focal lengths with 2 MP, 2/3-inch, F1.4 and C-Mount specifications. Kyptec Automation® specifically identifies moisture detection and material identification among the intended applications for these lenses, making the portfolio a strong optical platform to evaluate when an OEM is developing a dedicated SWIR moisture-inspection station.

Why Water Creates Useful Contrast in SWIR Images

Water molecules contain O–H bonds whose vibrational behaviour produces characteristic absorption bands in the near-infrared and short-wave infrared regions. Within a 900–1700 nm industrial SWIR system, the most important practical feature is the strong water absorption region near 1450 nm. Published optical research also identifies a weaker water-related feature around 1650 nm, while other water bands exist outside the upper limit of a conventional 1700 nm SWIR system.

For an industrial machine, this means that wet and dry regions can return different optical intensities even when both look similar to a human observer. In a typical reflectance arrangement, radiation illuminates the product, interacts with its surface and internal structure, and some portion is reflected toward the SWIR camera. Where water absorbs strongly, the returned signal around the water-sensitive wavelength can decrease. A wetter region may therefore appear darker than a drier region under appropriately selected illumination, although the actual contrast direction and magnitude must always be verified on the real material because surface scattering, composition and geometry also influence measured intensity.

This distinction is essential for OEMs. A SWIR image is not a universal “water map.” It is an optical measurement influenced by water absorption and the material surrounding the water.

Why 1450 nm Is Particularly Important Inside a 900–1700 nm System

The 900–1700 nm range is highly useful for industrial moisture inspection because it includes the prominent water-absorption region near 1450 nm while remaining within the spectral range of many industrial SWIR imaging architectures. Research consistently identifies approximately 1450 nm as a strong water-sensitive band, and industrial imaging references use it for applications such as moisture identification, fruit inspection and fill-level detection.

The practical implication is important: an OEM does not necessarily need to interpret every wavelength between 900 and 1700 nm equally. A complete spectral-development experiment may first determine which wavelength gives the strongest separation between the actual good and bad products. Once that relationship is understood, a production machine can sometimes be engineered around a smaller number of strategically selected spectral bands.

For moisture inspection, a water-sensitive wavelength can also be compared with a reference wavelength where the material response is less dominated by water. Comparing those two responses can be substantially more robust than classifying products from one absolute grayscale value.

Absolute Brightness Is Usually a Weak Moisture Metric

Suppose a production system classifies everything below grayscale value 80 as “wet.” That may work during a laboratory test and fail after installation.

Why? Because absolute image intensity changes when illumination output drifts, the product tilts, the working distance changes, the surface becomes more reflective, the lens aperture changes or the camera exposure is adjusted. A dark pixel does not automatically mean high moisture.

A stronger industrial design looks for relative spectral behaviour.

For example, let (I_w) represent intensity recorded near a water-sensitive wavelength and (I_r) represent intensity at a reference wavelength. An OEM may investigate normalized relationships such as:

Normalized moisture response = (Ir − Iw) / (Ir + Iw)

or another application-specific ratio derived during characterization.

The exact equation is not universal, but the engineering principle is valuable: normalization can reduce sensitivity to overall brightness variations and isolate the wavelength-dependent response that is more closely associated with moisture.

Moisture Detection Is Not Automatically Moisture Measurement

There is an important commercial distinction between detecting a moisture difference and reporting an absolute moisture percentage.

A SWIR machine may be very effective at answering questions such as: Is this product wetter than the approved reference? Is this region incompletely dried? Is there an abnormal wet patch? Is this product inside or outside the acceptable optical moisture signature?

Those are classification problems.

Reporting a value such as “12.4% moisture content” is considerably more demanding. It normally requires calibration against independently measured moisture values, representative samples across the expected product range, controlled imaging conditions and validation of the calibration model.

OEMs should therefore decide early whether the machine needs relative moisture detection, pass/fail classification, spatial moisture mapping or quantitative moisture estimation. Each requires a different validation strategy.

Surface Moisture and Internal Moisture May Produce Different Responses

SWIR radiation does not interact with every material to the same depth. Optical penetration depends on wavelength, absorption coefficient, scattering, surface structure and composition. Strong water absorption can increase moisture sensitivity while simultaneously reducing penetration depth.

Consequently, a dark water-sensitive region can sometimes be dominated by near-surface moisture rather than representing the complete bulk moisture content of a thick product. Scientific measurements of wet materials have also demonstrated that surface moisture can differ from bulk moisture and that reflectance does not always change linearly with total water content.

This is one of the most important limitations to understand before purchasing optics for a moisture-inspection system. A successful OEM qualification should compare SWIR response with the exact moisture condition the machine is expected to control rather than assuming that surface reflectance is automatically equivalent to total product moisture.

Reflectance Imaging for Surface and Distributed Moisture Differences

Reflectance geometry is often practical when illumination and camera are positioned on the same side of the product. SWIR radiation reaches the material and the camera measures the returned optical signal.

This architecture can work well for inspecting wet spots, drying uniformity, food surfaces, textiles, powders, webs and other products where moisture alters reflected SWIR intensity.

The challenge is that surface angle affects reflected intensity. A curved or irregular product can create brightness variation unrelated to moisture. The machine should therefore use carefully designed illumination geometry and, where possible, normalize water-sensitive information against another spectral response.

For wide product zones or compact conveyor layouts, the Kyptec Automation® KL-1410 12.5 mm SWIR Camera Lens provides a useful focal-length option within the 900–1700 nm portfolio. Its F1.4 aperture can also provide valuable light-collection margin where narrow spectral illumination reduces the optical energy reaching the sensor. The model is specified for 2 MP, 2/3-inch and C-Mount operation.

Transmission Imaging Can Be Powerful When the Product Geometry Allows It

In transmission inspection, illumination is placed behind the product and the camera measures radiation passing through it. Water-rich regions can attenuate water-sensitive wavelengths more strongly, potentially producing very high contrast.

Transmission geometry can be particularly valuable where the surrounding packaging or material transmits SWIR radiation while the water-containing region absorbs it. Industrial SWIR systems use this principle for tasks including liquid or fill-level visualization in circumstances where visible appearance gives inadequate information.

Transmission should not be assumed to work through every product. Material thickness, package composition and scattering may reduce the available signal substantially. Before designing the mechanical station, OEMs should test the actual product under the intended wavelength rather than assuming that a visually opaque material will necessarily be SWIR-transparent.

The SWIR Lens Must Preserve the Water-Related Signal

The camera and illumination may receive most of the attention in moisture-inspection development, but the lens is part of the spectral measurement path. If optical transmission becomes poor or inconsistent around the wavelength carrying the moisture information, material separation can deteriorate even if the camera itself is sensitive there.

A dedicated 900–1700 nm SWIR camera lens is therefore preferable to treating the lens as a generic mechanical accessory. The Kyptec Automation® SWIR range is specifically specified for 900–1700 nm operation and industrial SWIR imaging, providing a coherent optical platform for systems whose useful moisture information may occur around the strong 1450 nm absorption region.

Why F1.4 Can Matter in Moisture-Selective Imaging

Spectral moisture imaging can become light-limited. Narrow-wavelength illumination or optical filtering may reduce the total radiation reaching the sensor compared with broadband illumination.

The F1.4 maximum aperture specified across the Kyptec Automation® SWIR portfolio gives an OEM useful light-collection capability. This can help when the machine requires short exposure times, particularly on moving products.

However, operating fully open is not automatically optimum. Irregular products may require greater depth of field, forcing the iris to a smaller aperture. In that case, the lost optical signal may need to be recovered through illumination intensity, exposure time or another part of the imaging architecture.

The correct moisture-inspection aperture should therefore be established using actual production samples and the full expected height variation.

Focal Length Should Be Selected From the Moisture Region That Must Be Resolved

A common mistake is to choose focal length according to total product size without considering the smallest wet region that must be detected.

Suppose an inspection system covers 500 mm of conveyor width, but the smallest unacceptable moisture patch is only 3 mm. The camera must assign enough pixels to that 3 mm region for the optical moisture difference to remain distinguishable from noise, texture and illumination variation.

This is where focal length, field of view and moisture-detection sensitivity become connected.

For a controlled station where moderate coverage and higher object representation are preferable, the Kyptec Automation® KL-1412 25 mm SWIR Camera Lens provides a useful intermediate geometry within the current family. Its 25 mm focal length, F1.4 aperture, 2 MP class, 2/3-inch format and C-Mount architecture make it a logical lens to evaluate where a moisture-inspection system needs tighter framing than a wide-angle configuration.

A Moisture Patch Must Occupy Enough Pixels to Be Reliable

Detection is not based on whether a defect occupies one theoretical camera pixel. A robust classifier normally needs multiple pixels across the region so that real moisture contrast can be separated from sensor noise, optical blur, surface texture and interpolation.

An OEM should therefore convert the smallest moisture feature into object-side sampling.

If a 1600-pixel horizontal image covers 320 mm, sampling is approximately 5 pixels/mm. A 2 mm wet region would occupy about 10 pixels horizontally before accounting for optical resolution and contrast. If the same sensor covers 800 mm, the feature occupies only about 4 pixels.

This calculation is often more meaningful than comparing camera megapixels alone.

Moisture Contrast Should Be Tested Across the Entire Field

It is easy to obtain excellent moisture separation at the image center and assume the lens selection is complete. Production systems need consistent performance at the field edges as well.

During qualification, place representative wet and dry samples at the center, corners and intermediate locations. Measure the water-sensitive response at every position.

If classification threshold changes significantly across the field, investigate illumination uniformity, lens behaviour, mechanical angle and sensor response before attempting to compensate everything in software.

A stable optical system should minimize the amount of location-specific correction required downstream.

Drying Verification Is Often More Valuable Than Simple Wet/Dry Detection

Many industrial processes do not need to locate visible liquid. They need to know whether a material has reached an acceptable drying state.

Examples include processed food, pharmaceutical materials, coated products, textiles and industrial webs.

The value of SWIR in these applications is the possibility of detecting residual water-content differences before they become obvious visually. The machine can compare spatial regions, production batches or sequential process stages and identify material that remains outside the validated optical response window.

This turns the SWIR camera from a defect-viewing device into a process-control sensor.

Moisture Uniformity Can Matter More Than Average Moisture

Two products may have similar average water content but very different distributions.

One may be uniformly dry while another contains a localized wet region surrounded by very dry material.

A point measurement could potentially miss that difference, whereas imaging produces spatial information across the entire inspected area.

An industrial SWIR moisture inspection system can therefore be designed to evaluate not only mean response but also maximum local moisture signal, wet-region area, spatial variation and distribution uniformity.

For drying lines and coating processes, that spatial information can be more actionable than a single average value.

Food and Agricultural Products Require Product-Specific Calibration

Water is highly relevant to foods, fruit, vegetables, grains and other biological materials, and SWIR imaging is widely associated with agricultural and food inspection because moisture-related absorption creates useful contrast.

However, two food products with identical moisture percentages may not produce identical image intensities. Skin, sugar, fat, fibre, surface texture, thickness and internal scattering can all affect the optical response.

An OEM should therefore avoid developing one universal threshold and applying it across multiple product families.

Each material class should be characterized separately.

Pharmaceutical and Powder Inspection Requires Special Attention to Surface Geometry

Powders, granules and tablets can create significant scattering. Particle orientation and packing density may alter the amount of SWIR radiation reaching the camera even when moisture remains constant.

For this type of machine, repeated measurements should be taken across different surface conditions. A spectral ratio or normalized metric may provide greater robustness than an absolute grayscale threshold.

If the required inspection area is relatively small and the camera can operate with greater stand-off, a longer focal length can concentrate the sensor on the relevant region. The Kyptec Automation® KL-1414 35 mm SWIR Camera Lens, specified for 900–1700 nm, F1.4, 2 MP, 2/3-inch and C-Mount operation, is a strong option to evaluate for controlled moisture-analysis cells and similar tighter-field installations.

False Moisture Signals: What Can Mislead a SWIR Inspection System?

A dark region is not automatically water.

Several variables can imitate or modify moisture contrast: surface angle, shadows, product thickness, material composition, contamination, illumination non-uniformity, sensor drift, temperature-related changes and strongly absorbing non-water constituents.

Reliable system development therefore requires a confounder study. Collect samples representing wet material, dry material, acceptable variations, surface contamination, thickness extremes and expected production changes. If possible, deliberately introduce variables that should not trigger a moisture reject.

A machine that has only been tested on perfectly prepared wet and dry laboratory samples is not yet production-qualified.

Why Reference Samples Should Be Part of Commissioning

An OEM can improve long-term stability by keeping traceable reference samples or an optical reference target for routine verification.

During commissioning, the system should establish the expected response of known material conditions at the water-sensitive wavelength. Those results can become acceptance limits for later machines.

If production equipment is replicated across several factories, standardized optical references make it easier to determine whether two systems behave consistently.

This is particularly valuable when the aperture, exposure, working distance and illumination settings must be replicated between installations.

Temperature, Product Variation and Process Drift Need to Be Included in Validation

A prototype built in a laboratory is often tested under stable conditions. Production is less cooperative.

Products can arrive at different temperatures, surfaces can become dusty, illumination output can change gradually and the material recipe itself can vary.

A robust moisture-inspection algorithm should therefore be challenged across the expected operating envelope before the pass/fail threshold is frozen.

For an OEM, the correct question is not “Can SWIR distinguish these two samples?” The stronger question is: Can the complete 900–1700 nm imaging system maintain the required separation between acceptable and unacceptable moisture conditions after all legitimate production variation is introduced?

Building an OEM Acceptance Test for SWIR Moisture Detection

A strong acceptance protocol begins with independently characterized samples covering clearly dry, borderline and clearly wet conditions. Capture multiple images of every moisture class under expected production speed and object-position variation. Repeat the samples across the field of view and across the intended working-distance tolerance.

The OEM should then quantify class separation, repeatability and false-reject behaviour. The test should include exposure variation, illumination drift and representative product variation.

Only after those tests should a production threshold or classification model be released.

This method transforms SWIR moisture detection from an attractive demonstration into an engineered inspection process.

Why Kyptec Automation® Is a Strong Optical Platform for Moisture-Detection OEMs

Kyptec Automation® offers a dedicated family of SWIR camera lenses covering 900–1700 nm, a range that includes the important water-sensitive region near 1450 nm. The current portfolio provides five focal lengths from 8.5 mm through 50 mm, allowing machine builders to design broad conveyor views, intermediate inspection stations or tighter analysis zones while remaining within the same specialized SWIR lens family. The products share 2 MP, 2/3-inch, F1.4 and C-Mount positioning, giving OEMs a consistent foundation for developing multiple moisture-inspection geometries.

For an industrial buyer, this breadth matters because moisture contrast and inspection geometry must be solved together. The broadest field is not always correct, and neither is the longest focal length. Kyptec Automation® gives the designer several practical focal-length choices while keeping the spectral range aligned with SWIR moisture and material-inspection requirements.

Frequently Asked Questions About SWIR Camera Lenses for Industrial Moisture Detection

1. What wavelength is most useful for detecting water with a SWIR camera?

Within a conventional 900–1700 nm industrial SWIR range, approximately 1450 nm is particularly important because water has strong absorption in this region. A higher-water-content area can therefore return substantially less radiation at this wavelength in an appropriate reflectance configuration. Production systems should still test nearby and reference wavelengths on the actual product because the best class separation depends on the material as well as water itself.

2. Why can two products with different moisture levels look identical in visible light but different in SWIR?

Visible cameras primarily record differences produced by visible reflectance and colour. Water-content variation may not cause a strong visible colour change. Around water-sensitive SWIR wavelengths, however, absorption changes can substantially modify the returned optical signal. SWIR therefore provides material-dependent contrast based on spectral behaviour rather than relying only on visual appearance.

3. Can SWIR detect a small moisture difference or only obviously wet material?

Small differences can potentially be detected, but sensitivity is application-dependent. The result depends on the product, illumination wavelength, sensor performance, lens transmission, signal-to-noise ratio, exposure and calibration. OEM testing should include samples close to the actual production acceptance limit, not merely extreme wet and dry samples.

4. How do I know whether moisture should appear dark or bright in the final image?

Under reflectance illumination near a strong water-absorption wavelength, higher water content commonly produces reduced returned intensity and therefore a darker appearance. The final image can nevertheless be influenced by material scattering, thickness, illumination geometry and image-processing conventions, so the contrast direction should always be confirmed experimentally rather than assumed.

5. Can a SWIR moisture system distinguish water from a dark-coloured surface?

Often it can, because SWIR classification can use wavelength-dependent response rather than visible colour. A visibly dark surface does not necessarily remain dark at 1450 nm. A robust system should compare the response at a water-sensitive wavelength with one or more reference conditions instead of simply classifying every dark pixel as moisture.

6. Should I use one SWIR wavelength or two wavelengths for moisture inspection?

A single carefully selected wavelength can be enough for some controlled applications, but a water-sensitive wavelength plus a reference wavelength often provides a stronger basis for normalization. Comparing spectral responses can reduce errors from changes in overall brightness, product distance or reflectivity. The most appropriate architecture should be established through sample testing.

7. Can SWIR imaging detect uneven drying across a product?

Yes, this is one of the major advantages of imaging compared with a single-point moisture measurement. A camera can generate spatial information across the field, making it possible to identify localized wet zones, edge-to-centre differences and non-uniform drying. The inspection algorithm can then evaluate moisture distribution rather than only average response.

8. Can SWIR detect moisture through packaging?

It can in some applications when the packaging material transmits the selected SWIR wavelength sufficiently well. The packaging must be characterized before system design because materials that look similar in visible light can behave very differently in SWIR. Transmission loss, package thickness and printing can also affect the result.

9. How should I inspect a wet patch that is only a few millimetres wide?

Calculate the object-side pixel sampling before selecting the lens. The moisture patch should occupy enough pixels to remain distinguishable from optical blur, sensor noise and surface texture. If a wide field makes the defect too small in the image, the OEM may need a longer focal length, reduced FOV or higher-resolution imaging architecture.

10. Does higher moisture always produce a linearly darker SWIR image?

No. Industrial materials can exhibit nonlinear optical behaviour because absorption, scattering, surface coverage and penetration depth change with water content. Research on wet materials has demonstrated that reflectance is not necessarily a simple linear function of total moisture. Quantitative systems therefore need empirical calibration across the actual moisture range.

11. Can SWIR moisture imaging measure moisture below the surface?

Potentially, but the effective sampling depth depends strongly on wavelength and material properties. Stronger absorption can improve sensitivity to water while reducing how deeply radiation penetrates. An OEM should never assume that a surface SWIR image represents bulk moisture without validating the relationship against independently measured samples.

12. What is the best SWIR lens focal length for a moisture-detection conveyor?

There is no universal focal length. Broad conveyors or compact camera positions may favor shorter options, while localized moisture inspection can benefit from intermediate or longer focal lengths that devote more sensor pixels to the area of interest. Kyptec Automation® offers 8.5 mm through 50 mm SWIR options, allowing FOV to be matched to the real conveyor width and minimum wet-feature size rather than selecting focal length by application name alone.

13. Why does my SWIR moisture contrast change when the product moves across the image?

The cause may be non-uniform illumination, angle-dependent reflectance, lens behaviour, product geometry or sensor-response variation. Place identical reference samples at several image positions and compare their measured intensities. If the response changes with position rather than moisture, correct the optical uniformity before adjusting the classification threshold.

14. How do I prevent surface texture from being mistaken for moisture?

Use representative training and validation samples containing normal texture variations at controlled moisture levels. A reference wavelength or normalized spectral metric can also help separate broadband brightness variation from water-specific absorption. Illumination geometry should be optimized to reduce shadows and specular effects before relying on software compensation.

15. Can the same moisture threshold be used for different materials?

Usually this should not be assumed. Different materials have different baseline reflectance, scattering, thickness and chemical composition. Even if water absorption near 1450 nm is common to both, the measured camera intensity can differ substantially. Each material or product family should therefore receive its own characterization and, where necessary, its own calibration.

16. How should an OEM validate a SWIR moisture-inspection machine before delivery?

Validation should include known moisture classes, borderline samples, product-position variation, different surface conditions, expected temperature range, illumination drift and repeated measurements over time. Test samples should be placed throughout the FOV, and false positives from legitimate product variation should be measured. The machine should only be released once acceptable and unacceptable moisture states remain reliably separated across the full production envelope.

17. What specifications should I provide when selecting a SWIR lens for moisture detection?

Provide the required 900–1700 nm spectral operation, camera sensor dimensions, inspection width and height, working distance, smallest moisture region, product movement speed, height variation and mechanical space available around the inspection station. These details allow focal length and optical geometry to be selected from the real inspection requirement rather than from a generic application recommendation.

18. Why is Kyptec Automation® a strong choice for an OEM developing industrial SWIR moisture detection?

Kyptec Automation® provides a dedicated 900–1700 nm SWIR Camera Lens family spanning 8.5 mm, 12.5 mm, 25 mm, 35 mm and 50 mm focal lengths, with F1.4, 2 MP, 2/3-inch and C-Mount specifications across the current portfolio. Moisture detection is explicitly identified among the intended SWIR applications. This gives an OEM a focused optical family that can address multiple inspection widths and working-distance requirements while retaining the spectral coverage needed to work around the important 1450 nm water-absorption region.

Conclusion

The reason a 900–1700 nm SWIR camera lens can be so effective for industrial moisture detection is not that infrared simply makes water visible. The real mechanism is more useful and more precise: water interacts differently with specific SWIR wavelengths, producing absorption features that can create measurable contrast between regions containing different amounts or distributions of moisture. Around 1450 nm, that absorption becomes particularly strong within the operating range of conventional 900–1700 nm SWIR imaging systems.

Turning that physical principle into a reliable factory machine requires substantially more than pointing a SWIR camera at a wet sample. The OEM must choose an appropriate water-sensitive wavelength, establish a stable reference, control illumination and exposure, provide enough spatial resolution for the smallest wet region, account for surface geometry and material composition, distinguish surface response from bulk moisture where necessary, and validate the system against independently characterized production samples.

The Kyptec Automation® SWIR Camera Lens portfolio provides a strong optical foundation for that engineering process. Its dedicated 900–1700 nm coverage includes the critical 1450 nm water-sensitive region, while focal lengths from 8.5 mm to 50 mm allow the optical field to be adapted from broad conveyor inspection to tightly controlled moisture-analysis stations. The common 2 MP, 2/3-inch, F1.4 and C-Mount architecture also gives OEMs a coherent lens family when several inspection geometries must be evaluated within the same machine platform.

For industrial buyers, the strongest design principle is therefore to treat SWIR moisture inspection as a calibrated spectral measurement, not simply as a darker-versus-brighter image test. When the lens, wavelength, illumination, sensor, field of view and reference methodology are engineered together, SWIR can reveal water-content differences that remain difficult or impossible to classify reliably from visible appearance alone—and that is where a purpose-built Kyptec Automation® SWIR Camera Lens becomes most valuable in an automated moisture-control system.