SWIR Reflectance vs Transmission Imaging: How to Choose the Right Optical Geometry for Industrial Inspection

A SWIR inspection system can use exactly the same camera and lens yet produce completely different information depending on where the illumination is placed. This is one of the most important—and frequently underestimated—decisions in short-wave infrared machine vision. In SWIR reflectance imaging, illumination and camera generally observe the object from the same side, and the system analyzes wavelength-dependent radiation returned from the surface and near-surface structure. In SWIR transmission imaging, illumination is positioned behind the inspected material so the camera measures radiation that survives passage through the object. The first geometry is usually stronger for material identification, surface composition, contamination and moisture-related contrast; the second can be considerably stronger for selected internal features, hidden contents, inclusions, fill conditions and defects inside materials that transmit enough SWIR radiation. Industrial imaging literature similarly distinguishes reflected and transmitted infrared inspection according to whether useful information is returned from or travels through the object.

Choosing between these geometries should happen before the SWIR camera lens focal length is finalized. A machine builder who first selects a lens and later decides where the illumination will be positioned can easily create a mechanically awkward system with insufficient optical signal, uncontrolled reflections or an unusable working distance. The correct sequence is to identify the defect or material property, determine where the useful SWIR contrast originates, choose reflectance, transmission or a qualified combination of both, and then calculate field of view, working distance and focal length.

The Kyptec Automation® SWIR Camera Lens collection provides five focal lengths—8.5 mm, 12.5 mm, 25 mm, 35 mm and 50 mm—within a dedicated 900–1700 nm optical platform. Current product specifications include 2 MP resolution class, 2/3-inch sensor format, F1.4 aperture and C-Mount. Kyptec Automation® positions the range for material identification, moisture detection, food processing, pharmaceutical, electronics and other industrial inspection applications where visible imaging is insufficient. This makes the portfolio especially useful for OEMs that need to optimize optical geometry around the actual physics of the inspection rather than force every SWIR application into one standard camera arrangement.

What SWIR Reflectance Imaging Actually Measures

In a reflectance configuration, illumination reaches the object and some fraction of that radiation is absorbed, some scattered internally and some returned toward the camera. The recorded intensity therefore depends on material absorption, surface texture, illumination angle, wavelength, object geometry and the optical characteristics of the imaging system.

A simplified reflectance relationship can be expressed as:

R(λ) = I reflected(λ) / I incident(λ)

where the response varies with wavelength (λ).

The important point is that the camera does not simply measure whether an object is “bright” or “dark.” It measures how efficiently a material returns radiation at different SWIR wavelengths. Two visually identical products can therefore produce different reflectance profiles if their chemical composition differs.

This is one reason 900–1700 nm spectral imaging is widely used for material classification, contamination detection and compositional inspection. Industrial studies in this wavelength region show that spectral differences can separate materials that are difficult to discriminate from visible appearance alone.

What SWIR Transmission Imaging Measures

Transmission inspection analyzes the radiation that passes through an object or material stack.

A simplified relationship is:

T(λ) = I transmitted(λ) / I incident(λ)

The transmitted signal depends on the absorption and scattering accumulated throughout the optical path.

This gives transmission imaging an important advantage: it can contain information from material inside the object rather than predominantly from its illuminated surface.

When an inclusion, internal structure, product fill or defect attenuates SWIR differently from the surrounding material, the transmitted image can reveal it with high contrast. Reviews of hyperspectral inspection describe transmittance imaging as increasingly useful for evaluating internal features in appropriate materials and products.

However, the geometry requires access for illumination behind the object and enough SWIR transmission through the complete material thickness.

Reflectance and Transmission Are Not Competing Technologies

A common mistake is asking which geometry is universally better.

Neither is better in isolation.

They answer different physical questions.

Reflectance is usually preferable when the target information exists in the surface or near-surface optical response, or when the object is too thick or opaque for useful transmission.

Transmission becomes attractive when the inspection target lies inside a material that permits sufficient SWIR radiation to pass through it.

The correct choice is therefore:

Where is the information physically located, and how can SWIR radiation interact with that location most effectively?

This one question can prevent substantial trial-and-error during machine development.

Choose Reflectance for Surface and Material Classification

Reflectance geometry is usually the natural starting point for SWIR material identification.

Examples include separating different polymers, distinguishing grain or seed types, classifying textile fibers, identifying contamination on food, evaluating surface moisture differences or verifying whether one material class has been substituted for another.

These applications depend primarily on wavelength-dependent material response rather than imaging through the complete thickness of the object.

Reflectance is also mechanically simpler because both illumination and camera can remain on one side of the production line.

For conveyors, sorting systems and large web inspection, this can significantly simplify guarding, cleaning and maintenance.

Choose Transmission When the Defect Is Hidden Inside a Transmitting Material

Transmission becomes especially valuable when the feature of interest is inside a product or behind another material.

Potential examples include:

hidden content inside selected packaging;

inclusions inside SWIR-transmitting products;

internal fill boundaries;

selected subsurface food defects;

features behind silicon;

and defects within partially transmitting polymeric structures.

Transmission imaging has been widely investigated for internal quality assessment because the radiation reaching the camera has interacted with a greater portion of the object volume than in conventional surface reflectance.

The limitation is equally important: if the product absorbs nearly all incident SWIR, there will be too little transmitted signal to form a useful image.

Optical Density Determines Whether Transmission Is Practical

A material does not need to be perfectly transparent for transmission imaging to work.

The relevant question is whether sufficient radiation remains after travelling through the object.

Using the Beer-Lambert concept in simplified form:

I = I₀e⁻ᵅˣ

where (I₀) is incident radiation, (α) represents effective attenuation and (x) is optical path length.

As thickness increases, transmitted intensity can fall rapidly.

This explains why a 1 mm polymer sheet may transmit enough SWIR while the same material at 15 mm thickness becomes difficult to inspect.

Machine builders should therefore test the maximum production thickness, not only the easiest sample.

Object Thickness Changes the Reflectance-vs-Transmission Decision

Thin films, wafers, translucent polymers and selected food products can be strong transmission candidates.

Thick, highly scattering or strongly absorbing products are more likely to favor reflectance.

This is particularly important for products with variable thickness.

If a system works only on 5 mm samples but production ranges from 5 to 20 mm, the transmitted signal can vary so much that classification becomes unstable.

Thickness should therefore be treated as a primary inspection parameter alongside wavelength, field of view and working distance.

Water Can Strengthen or Complicate Both Geometries

Water produces significant absorption features in the SWIR region, which is why 900–1700 nm imaging is useful for moisture-related inspection.

In reflectance imaging, water-rich areas may change the wavelength-dependent returned signal, making moisture variation visible.

In transmission, water can strongly reduce the amount of radiation passing through the product, which can create very clear liquid or moisture boundaries.

However, water can also become a nuisance variable. Surface condensation may alter reflectance even when the underlying material is acceptable. In transmission, different thicknesses of a water-containing product can produce different attenuation without representing a defect.

The machine must therefore distinguish composition variation from path-length variation.

Why Reflectance Is Common in Material Sorting

Material sorting machines need simple mechanical access, high throughput and stable views of large product streams. Reflectance geometry usually satisfies these requirements more naturally.

Illumination can be installed around or beside the camera, while objects move along a conveyor or chute.

The system then classifies products according to their returned SWIR spectral signature.

This approach is especially useful for plastics, agricultural products, textiles and foreign-material separation.

Research using 900–1700 nm imaging continues to demonstrate strong material and contamination classification capability in production-oriented inspection scenarios.

Why Transmission Can Reveal Hidden Defects More Clearly

Consider a foreign particle embedded inside a material.

In reflectance mode, radiation may interact strongly with the upper surface before reaching the inclusion. If the overlying material is scattering or absorbing, the particle contributes little to the returned signal.

In transmission mode, the radiation must pass through the product before reaching the camera. If the inclusion attenuates the beam differently from the surrounding material, its presence can become much more obvious.

This is why transmission can be superior for internal inclusions, selected subsurface defects and concealed contents where optical penetration is sufficient.

Surface Texture Is a Bigger Challenge in Reflectance Imaging

Roughness, curvature, gloss and orientation strongly influence reflected intensity.

A glossy product can create specular highlights. A rough surface can scatter light diffusely. A curved object can appear brighter in one region and darker in another even though its material is uniform.

If a classifier uses absolute grayscale thresholds, these geometric effects can be mistaken for material differences.

Reflectance systems should therefore control illumination angle carefully and often benefit from normalized multi-wavelength features rather than one raw intensity value.

Transmission Reduces Some Surface Effects but Introduces Thickness Effects

Transmission can reduce sensitivity to certain surface-reflection effects because the camera is primarily measuring radiation passing through the product.

But the geometry introduces a different variable: optical path length.

A curved container, thick product region or overlapping objects will transmit less radiation even if material composition is unchanged.

The choice is therefore not between an imperfect and perfect geometry.

It is between different sources of measurement variation, and the better system is the one whose nuisance variables can be controlled more effectively.

Reflection Angle Should Be Designed, Not Left to Chance

In a reflectance station, camera and light should not automatically be placed directly beside each other.

For glossy materials, near-coaxial illumination can send strong specular reflection into the lens and reduce useful material contrast.

Moving illumination off-axis can suppress glare while preserving diffuse material response.

For highly textured objects, broader diffuse illumination may provide more consistent spectral measurement.

The best angle should be selected from actual contrast measurements using known-good and known-bad samples.

Transmission Illumination Must Be Uniform Across the Entire Field

Backlighting for SWIR transmission should be spatially uniform.

Suppose the center of a light source is 20% brighter than its edges. The camera can then interpret identical products differently depending on position.

A reference correction can compensate for some illumination non-uniformity, but good optical design should minimize the variation before software normalization.

Flat-field images acquired without the product can be useful for evaluating whether the illumination-plus-lens combination produces sufficient uniformity across the sensor.

Kyptec Automation® KL-1408 for Broad Reflectance Inspection

For wide conveyors, large sorting areas or broad material webs, the Kyptec Automation® KL-1408 8.5 MM SWIR Camera Lens provides the widest focal-length geometry in the current Kyptec Automation® SWIR portfolio. Its verified specifications include 8.5 mm focal length, 900–1700 nm operation, F1.4 aperture, 2 MP resolution, 2/3-inch format and C-Mount.

This geometry is particularly useful to evaluate for broad reflectance systems where the camera must observe a large conveyor or material surface from limited height.

The trade-off is spatial sampling. If the required defect is small, excessive FOV can reduce the number of pixels available to identify it.

Kyptec Automation® KL-1410 for Balanced Reflectance or Transmission Fields

The Kyptec Automation® KL-1410 12.5 MM SWIR Camera Lens provides a useful intermediate wide-field geometry with verified 900–1700 nm, F1.4, 2 MP, 2/3-inch and C-Mount specifications.

This focal length can suit reflectance systems that need somewhat stronger object sampling than an 8.5 mm configuration while retaining broad production coverage.

It can also work well in transmission architectures where several packages, pieces or inspection regions must remain inside one backlit field.

The important point is that focal length does not determine whether the system is reflectance or transmission. It determines how much of the chosen optical geometry is mapped onto the sensor.

Kyptec Automation® KL-1412 for Controlled Material and Internal Inspection

For narrower FOV inspection, the Kyptec Automation® KL-1412 25 MM SWIR Camera Lens allows more of the camera's sensor area to be dedicated to a smaller target.

This can be valuable for controlled reflectance material analysis, localized contamination inspection or transmission imaging through individual packages and components.

A 25 mm configuration is often useful when the machine has moved beyond broad sorting and now needs stronger representation of one specific region.

When 35 mm and 50 mm Become More Useful

Longer focal lengths become relevant where the inspection field is smaller, the available working distance is greater or the target should occupy more of the image.

The Kyptec Automation® KL-1414 35 MM SWIR Camera Lens and Kyptec Automation® KL-1416 50 MM SWIR Camera Lens can therefore be evaluated for tighter component, package, material or transmitted-feature inspection.

Longer focal length should not be mistaken for automatically better sensitivity. If the inspection field becomes too narrow for normal product position tolerance, production reliability can become worse.

Reflection and Transmission Can Be Compared Quantitatively

Rather than deciding visually, an OEM can compare the two geometries using a contrast metric.

For two material states A and B:

Normalized contrast = |IA − IB| / (IA + IB)

The measurement can be repeated at each candidate wavelength in both reflectance and transmission.

Suppose reflectance produces normalized contrast of 0.12 while transmission produces 0.42 for the same hidden defect. Transmission offers a substantially stronger starting point.

If the opposite occurs for a surface contamination problem, reflectance is the better architecture.

This converts geometry selection from opinion into measurable engineering evidence.

Signal-to-Noise Ratio Matters as Much as Contrast

A geometry can produce a large percentage difference but still be unusable if very little signal reaches the sensor.

For example, two transmitted states might measure 2 and 4 arbitrary units. That is a 2:1 difference but may remain dominated by sensor noise.

Another configuration producing 800 and 1000 units has lower relative contrast but far stronger measurement stability.

OEM selection should therefore consider both contrast and signal-to-noise ratio.

The F1.4 capability across the Kyptec Automation® SWIR range provides valuable optical throughput where available SWIR signal is limited, particularly in transmission configurations where the product attenuates illumination significantly.

FOV Calculation Does Not Change Between the Two Geometries

The geometric relationship between sensor, focal length, working distance and field of view remains fundamental whether the system is operating in reflection or transmission.

A simplified object-width estimate is:

FOV ≈ Sensor Width × Working Distance / Focal Length

For the same sensor and working distance, increasing focal length narrows the field.

What changes is the mechanical definition of working distance and the required space for illumination.

A transmission machine needs room behind the product for the light source and often requires better product-plane control. A reflectance machine may have more freedom behind the object but require careful management of light-to-camera angle.

Transmission Is Often Harder to Integrate Mechanically

Reflection usually allows camera, lens and illumination to remain on the same side of the process.

Transmission requires an optical path through the production line.

Conveyor belts, fixtures, rollers, carriers or machine frames can block the rear illumination. A SWIR-transmitting belt or open inspection gap may be required.

This mechanical requirement should be evaluated early.

An application that is optically excellent in transmission but impossible to integrate into the production machine may ultimately need a reflectance solution with slightly lower contrast.

Reflectance Is Often Better for Thick or Opaque Objects

If nearly all incident SWIR is absorbed or scattered before reaching the far side of an object, transmission becomes impractical.

Reflectance can still provide material information from the surface and near-surface region.

This makes it particularly valuable for thick bulk materials, many agricultural products, textiles, polymer pieces and objects whose internal structure is not the inspection target.

The geometry should always match the depth from which useful information must be obtained.

Hybrid Development Can Identify the Best Production Architecture

During feasibility development, it can be useful to test the same samples in reflectance and transmission before committing to one geometry.

This does not mean the final production machine must use both.

The purpose is to learn which optical interaction creates the largest defect-to-good separation.

A research setup can therefore acquire reflected and transmitted spectral datasets from the same samples. The production design can then retain only the geometry that offers the required contrast with acceptable mechanical complexity.

This is often more efficient than trying to optimize a poorly chosen architecture after the machine is already built.

Why Kyptec Automation® Is a Strong Platform for SWIR Geometry Optimization

The Kyptec Automation® SWIR Camera Lens collection gives machine builders five focal-length choices within one dedicated 900–1700 nm lens family. The verified live product range spans 8.5 mm, 12.5 mm, 25 mm, 35 mm and 50 mm while maintaining F1.4 aperture, 2 MP resolution class, 2/3-inch format and C-Mount.

This is particularly useful for reflectance-versus-transmission development because an OEM can separate two decisions that are often confused. First choose how SWIR radiation should interact with the product. Then choose which focal length maps the required inspection area onto the sensor.

Kyptec Automation® provides a strong optical platform for that engineering process because broad reflectance fields, multi-object transmission stations, intermediate material-analysis cells and tighter hidden-feature systems can all be evaluated within the same dedicated SWIR Camera Lens family.

Frequently Asked Questions About SWIR Reflectance and Transmission Imaging

1. What is the main difference between SWIR reflectance and transmission imaging?

Reflectance imaging measures SWIR radiation returned from an illuminated object, while transmission imaging measures radiation that passes through the object before reaching the camera. Reflectance is usually more sensitive to surface and material properties; transmission can be more effective for selected internal features and hidden contents when sufficient radiation passes through the material.

2. Which SWIR geometry is better for material identification?

Reflectance is commonly the stronger starting point because different materials can produce distinct wavelength-dependent reflected spectra without requiring radiation to travel through the complete object. This makes it practical for polymer sorting, textiles, agricultural products and contamination classification.

3. Which geometry is better for internal defect inspection?

Transmission is often preferable when the defect lies inside a material that is sufficiently SWIR-transmissive. The radiation passing through the internal structure can produce stronger inclusion or defect contrast than surface reflectance. The actual material thickness and wavelength must still be tested.

4. Can a SWIR inspection system use both reflectance and transmission?

Yes. A development platform can use both geometries to compare contrast, and some specialized systems may retain both. For most production machines, however, one geometry should be chosen if it provides sufficient information because a simpler optical architecture is easier to calibrate and maintain.

5. How do I know whether my product transmits enough SWIR light?

Measure transmitted intensity through the actual production sample across the required wavelength range and compare it with an open-beam reference. Perform the measurement at maximum expected material thickness. A visually transparent or opaque appearance is not an adequate predictor of SWIR transmission.

6. Why can a material look uniform in reflectance but reveal defects in transmission?

A surface can have nearly constant reflectance while the internal structure varies. Transmission forces the radiation to interact with more of the object's thickness, so an inclusion, void, liquid region or hidden content can alter the amount of radiation reaching the camera even when the external surface looks identical.

7. Why can reflectance imaging produce false material differences on glossy products?

Specular reflection depends strongly on surface angle. A glossy curved region may return much more radiation toward the lens than an identical material tilted slightly differently. Off-axis or diffuse illumination and multi-wavelength normalization can reduce these geometry-driven variations.

8. Does transmission imaging eliminate problems caused by object geometry?

No. It reduces some reflection-related effects but becomes sensitive to optical path length. Thicker or overlapping regions transmit less radiation even when the material is identical. Product thickness and orientation therefore need to be controlled or modeled.

9. Is reflectance better for moisture detection?

Reflectance can be very effective for surface and near-surface moisture differences because water alters SWIR absorption. Transmission can be even stronger for certain fill or bulk-water measurements when the sample transmits enough radiation. The correct geometry depends on whether the target moisture is superficial or distributed through the object.

10. Is transmission imaging suitable for a conveyor system?

Yes, provided illumination can be positioned behind the inspected material and the conveyor or support does not block the SWIR path. This may require an open inspection gap or appropriately transmitting support arrangement. Mechanical integration should be checked before the optics are finalized.

11. Which Kyptec Automation® SWIR lens is suitable for a wide reflectance field?

The Kyptec Automation® KL-1408 8.5 MM SWIR Camera Lens is the widest focal-length option in the current portfolio and can be evaluated for large conveyor or material-web coverage. Final suitability depends on working distance and smallest feature size.

12. Which focal length is useful for an individual object in transmission?

An intermediate focal length such as the Kyptec Automation® KL-1412 25 MM SWIR Camera Lens can be useful where an individual package, component or material region should occupy more of the sensor. The required FOV should still be calculated from actual object dimensions.

13. Does opening the SWIR lens to F1.4 improve transmission inspection?

It can increase the optical signal reaching the camera, which is especially valuable when the object attenuates much of the illumination. However, using the maximum aperture reduces depth-of-field margin. The production aperture should be chosen from signal-to-noise ratio, focus tolerance and feature contrast together.

14. Should the same exposure be used for reflectance and transmission tests?

Usually not. Transmission often produces less signal because radiation passes through the product, so exposure or illumination may need to increase. Comparisons should be made after each geometry has been adjusted to operate within a useful sensor range without saturation.

15. What is the best way to compare SWIR reflectance and transmission experimentally?

Use the same known-good and known-defective samples, test several relevant wavelengths, normalize the illumination and measure defect-to-good contrast plus signal-to-noise ratio for each geometry. Also evaluate production variables such as thickness, position and speed before selecting the final architecture.

16. Can transmission imaging work through several material layers?

Yes, if the combined stack still transmits sufficient SWIR radiation. However, each additional layer contributes absorption, scattering and reflection. The complete production stack should be tested because individual materials that transmit well separately may provide insufficient signal when combined.

17. Why can one wavelength work better in reflection while another works better in transmission?

The optical path is different. In reflectance, the measured signal depends strongly on surface and shallow material interactions. In transmission, radiation experiences the full material thickness. Absorption bands that provide useful reflectance contrast may attenuate too strongly in transmission, while weaker absorption bands can provide better transmitted signal.

18. What information should an OEM provide before selecting a SWIR lens and geometry?

Provide the target material, defect type, whether the feature is on the surface or internal, material thickness, expected SWIR transmission, sensor dimensions, inspection width, working-distance constraints, smallest required feature, production speed and available illumination positions. These details allow geometry and focal length to be selected as one optical system.

19. Why is Kyptec Automation® a strong choice for reflectance and transmission SWIR inspection development?

Kyptec Automation® provides a dedicated SWIR Camera Lens portfolio spanning 8.5 mm, 12.5 mm, 25 mm, 35 mm and 50 mm within a consistent 900–1700 nm, F1.4, 2 MP, 2/3-inch and C-Mount architecture. This gives OEMs practical flexibility to optimize reflectance or transmission geometry without changing to an unrelated optical family every time FOV or working-distance requirements change.

Conclusion

Choosing between SWIR reflectance and transmission imaging is fundamentally a decision about where useful optical information is generated. Reflectance measures radiation returned from the product and is particularly powerful for material identification, surface contamination, moisture differences, fiber or polymer classification and applications where rear optical access is impractical. Transmission measures radiation that survives passage through the object and can reveal selected internal features, concealed contents, inclusions, fill conditions and defects that may remain weak or invisible in surface imaging. Industrial imaging research confirms that both reflected and transmitted infrared geometries have important roles, with the correct configuration depending on the location and optical properties of the target feature.

The strongest system design does not begin by asking which geometry sounds more advanced. It begins by characterizing the actual object. Determine whether the target is at the surface or inside the product. Measure how the material absorbs and transmits SWIR across the required wavelength range. Test minimum and maximum material thickness. Compare normalized defect contrast and signal-to-noise ratio in reflection and transmission. Then evaluate whether the preferred geometry can be integrated mechanically into the production machine.

Only after that decision should focal length be finalized. The Kyptec Automation® SWIR Camera Lens collection provides a particularly useful platform for this design process. The Kyptec Automation® KL-1408 8.5 MM SWIR Camera Lens can support broad reflectance or transmission fields, while the Kyptec Automation® KL-1410 12.5 MM SWIR Camera Lens provides a useful balance between coverage and object sampling. Intermediate and longer focal lengths can progressively concentrate the camera's sensor on smaller inspection regions. Across this range, the common 900–1700 nm, F1.4, 2 MP, 2/3-inch and C-Mount architecture gives OEMs a coherent specialized platform for developing several SWIR inspection geometries.

For machine builders and industrial buyers, the most useful design principle is therefore simple: choose the light path before choosing the lens. If the useful information returns from the object, engineer reflectance. If the information is revealed by radiation travelling through the object, evaluate transmission. If the answer is uncertain, measure both using representative good and defective production samples. Once the correct contrast mechanism is proven, Kyptec Automation® SWIR Camera Lenses can provide the focal-length flexibility required to convert that optical principle into a practical, repeatable industrial inspection system.