SWIR Camera Lens Distortion Explained: How TV Distortion Affects Inspection Accuracy, Edge Geometry and Machine Calibration

A SWIR camera lens can produce a sharp, high-contrast image and still introduce geometric error. This distinction matters whenever an industrial inspection system does more than classify materials or detect obvious defects. If the machine measures dimensions, calculates object position, determines edge location, compares geometry across the field of view, guides a mechanism, maps defect coordinates or relies on calibrated image-to-world conversion, lens distortion becomes part of the measurement error budget. In these applications, a straight physical edge should ideally remain straight in the image and an equal physical displacement should correspond to a predictable pixel displacement throughout the usable sensor area. Real optical systems depart from that ideal to some degree, especially toward the edge of the field.

For SWIR inspection, this issue deserves specific attention because a machine may combine material-sensitive 900–1700 nm imaging with dimensional or positional analysis. A system inspecting semiconductor structures, electronic components, packaging, textiles, food products or industrial materials may first use SWIR contrast to reveal the feature and then use its image coordinates to decide where the feature is, how large it is or whether it lies inside an acceptable region. At that point, TV distortion, edge geometry and calibration quality directly influence inspection accuracy.

The Kyptec Automation® SWIR Camera Lens collection currently includes 8.5 mm, 12.5 mm, 25 mm, 35 mm and 50 mm focal-length models designed for 900–1700 nm imaging, with 2 MP resolution class, 2/3-inch format, F1.4 aperture and C-Mount across the family. Kyptec Automation® describes the range as engineered for high SWIR transmission, contrast and low distortion for industrial machine vision applications. The important buyer question, however, is not simply whether a lens is described as low distortion. It is how much distortion exists, where it appears in the field, whether the application is sensitive to it, and whether calibration can reduce the resulting coordinate error sufficiently.

What Is TV Distortion in a SWIR Camera Lens?

TV distortion is a practical optical specification describing how much the image geometry departs from ideal rectilinear projection. In a perfect lens, straight lines in the object plane remain straight and their positions scale predictably from the center toward the edge of the image. In a real lens, radial magnification can change with image height, causing straight lines to bend inward or outward.

A simplified distortion calculation is often expressed as:

Distortion (%) = (Actual image height − Ideal image height) / Ideal image height × 100

The exact industry calculation convention can vary, so an OEM should compare lens specifications using consistent definitions. The most useful interpretation is that distortion represents a geometric mapping error, not a sharpness measurement.

A lens can resolve very fine detail but still bend geometry. Conversely, a lens can have extremely low distortion but insufficient resolution for the smallest defect. Resolution and distortion must therefore be evaluated separately.

Barrel Distortion and Pincushion Distortion Explained

The two most familiar radial distortion patterns are barrel distortion and pincushion distortion.

With barrel distortion, image magnification decreases progressively toward the edge. A rectangular grid appears to bulge outward, as though wrapped around a barrel. Straight lines near the image boundary bow away from the image center.

With pincushion distortion, magnification increases toward the edge. Straight grid lines bend inward, creating the appearance of a pinched image.

The sign convention used in a datasheet indicates the direction according to that manufacturer's definition. For machine vision engineering, the magnitude and repeatability are usually more important than the sign alone. A known, stable distortion can often be calibrated. An unstable mapping that changes with focus, working distance, temperature or mechanical assembly is substantially more difficult to compensate.

Why Distortion Is Often Least Visible at the Image Center

Radial distortion generally increases with distance from the optical axis. The center of the image can therefore look geometrically accurate while errors become progressively larger toward the corners.

This is why an inspection prototype can appear excellent when the reference object is centered but produce positional or dimensional errors after the same object moves toward an edge.

If the production line allows parts to arrive anywhere across the field, calibration should cover the complete usable image area, not only the central region.

This is also why low distortion is particularly valuable when a SWIR machine uses almost the full 2/3-inch sensor rather than restricting analysis to a small central region.

Why TV Distortion Matters Less for Some SWIR Applications

Not every SWIR application requires tight geometric accuracy.

Suppose a recycling machine only needs to determine whether a plastic fragment belongs to material class A or material class B. If classification depends on average spectral response over the object, a small amount of geometric distortion may have little effect.

Similarly, a moisture inspection system identifying whether a large region is wet or dry may tolerate moderate geometric deformation if the moisture classification remains stable.

This changes immediately when the same machine must report that the wet region starts exactly 18 mm from an edge, calculate its physical area, position a robot to remove it or compare its dimension against a tolerance.

Distortion importance should therefore be determined from the output required from the image, not from the industry name of the application.

Classification, Detection and Measurement Have Different Distortion Sensitivity

It is useful to separate three SWIR machine-vision tasks.

Classification asks what material or condition is present. Distortion is often secondary unless it changes the region being sampled.

Detection asks whether a defect exists and possibly where it is. Distortion becomes more important because feature coordinates must remain reliable.

Measurement asks how large, far apart, aligned or geometrically correct the features are. Distortion can become a major source of systematic error.

An OEM should therefore not specify one universal distortion tolerance for every SWIR machine. A material sorter and a precision calibrated inspection station can legitimately require different optical acceptance limits.

A Small Distortion Percentage Can Become a Large Physical Error

A percentage can sound insignificant until it is converted into object-space units.

Suppose the horizontal field of view is 400 mm and an edge-region mapping error corresponds roughly to 0.5% of the relevant dimension. A first-order estimate is:

400 mm × 0.005 = 2 mm

A two-millimetre error may be irrelevant for identifying a 100 mm food product, but unacceptable for locating a 0.5 mm component feature.

This example should not be treated as a universal direct conversion from TV distortion to local measurement error, because distortion varies across the field and calibration geometry matters. It illustrates the correct engineering principle: convert optical distortion into the physical units of the production tolerance before deciding whether it matters.

Kyptec Automation® KL-1408: Wide FOV Requires Greater Attention to Edge Geometry

The Kyptec Automation® KL-1408 8.5 MM SWIR Camera Lens provides the widest geometry in the current Kyptec Automation® SWIR portfolio. Its official datasheet specifies 8.5 mm focal length, 900–1700 nm designed wavelength, 2 MP, 2/3-inch image format, F1.4 and TV distortion of -1.1%.

That specification is particularly important because wide-angle lenses naturally deserve careful edge-of-field evaluation. The Kyptec Automation® KL-1408 can be very useful when a broad conveyor, material web or large inspection area must fit within the available working distance, but an OEM performing coordinate-based measurement should characterize residual error across the complete field rather than relying on central calibration alone.

For broad presence detection or material classification, -1.1% may be entirely manageable depending on the application. For dimensional metrology, calibration and an application-specific tolerance study become much more important.

Kyptec Automation® KL-1410: Lower Distortion for Wide-to-Medium Coverage

The Kyptec Automation® KL-1410 12.5 MM SWIR Camera Lens provides a useful transition between wide coverage and tighter geometric control. Kyptec Automation® specifies the model for 900–1700 nm, F1.4, 2 MP, 2/3-inch format and C-Mount. Its detailed optical specification gives TV distortion of approximately -0.23%, substantially lower in magnitude than the 8.5 mm model.

For OEMs that do not require the maximum width of an 8.5 mm lens, this can be a valuable design trade-off. Reducing field of view slightly can increase object sampling while also placing the system on a lens geometry with much lower published distortion.

That does not automatically make 12.5 mm the better lens. The correct choice depends on required FOV, available working distance and the measurement error budget.

The Kyptec Automation® KL-1412 25 MM Option for Controlled Geometric Inspection

The Kyptec Automation® KL-1412 25 MM SWIR Camera Lens is especially relevant where the machine can use a narrower field and allocate more sensor area to the inspection target. The model's detailed optical specification gives TV distortion of approximately -0.25%, while the live product architecture maintains 900–1700 nm operation, F1.4, 2 MP, 2/3-inch format and C-Mount.

A 25 mm focal length can be a strong practical choice for component-level SWIR inspection, controlled material analysis and localized dimensional verification where a very wide field is unnecessary.

The key point is that lower FOV and low distortion work together differently from resolution alone. More pixels across the object improve sampling, while controlled geometric distortion reduces systematic coordinate deformation. Both contribute to measurement performance.

Why Kyptec Automation® KL-1414 Is Particularly Interesting for Low-Distortion SWIR Inspection

The Kyptec Automation® KL-1414 35 MM SWIR Camera Lens is a particularly strong model to evaluate for geometry-sensitive SWIR applications because its detailed specification lists TV distortion of approximately -0.05%.

That is a very low published distortion magnitude within the current SWIR family.

The 35 mm focal length produces a narrower field than the 8.5 mm, 12.5 mm and 25 mm models, so it is best suited where machine space allows the required working distance and where the target region can fit comfortably inside that field.

For applications such as precise feature localization, calibrated inspection of electronics, geometry-sensitive semiconductor structures, controlled dimensional analysis or measurement of material regions revealed by SWIR, this model is especially relevant to evaluate.

Its advantage should still be verified in the complete camera-lens-working-distance configuration, because final machine accuracy depends on more than datasheet distortion alone.

Kyptec Automation® KL-1416 for Narrow Fields With Low Published Distortion

The Kyptec Automation® KL-1416 50 MM SWIR Camera Lens provides the longest focal length in the current Kyptec Automation® SWIR range. Its detailed optical specification gives TV distortion of approximately -0.08%, while its common platform includes 900–1700 nm operation, F1.4, 2 MP, 2/3-inch format and C-Mount.

This makes it useful to evaluate where a relatively small region must be inspected from greater stand-off or where geometric fidelity is important and the required FOV is narrow.

As with any longer focal length, the machine must retain enough positioning margin. Extremely tight framing can improve sensor utilization but increase the risk that normal part displacement moves the target outside the calibrated region.

Distortion and Field of View Must Be Considered Together

A lower-distortion longer focal length is not automatically preferable if the machine cannot physically achieve the required FOV.

Suppose an inspection requires 300 mm horizontal coverage at the available camera height. A 35 mm lens may provide excellent geometric characteristics but an insufficient field. The OEM then has three options: increase working distance, use a shorter focal length or redesign the number and placement of cameras.

This is why the decision sequence should be:

required FOV → allowable working distance → candidate focal length → distortion review → calibration feasibility → final acceptance test.

Choosing purely from the smallest distortion percentage can produce a lens that is geometrically attractive on paper but unusable mechanically.

Distortion Is Not the Same as Perspective Error

This distinction is essential in machine calibration.

Lens distortion is caused by the optical mapping of the lens. Perspective error is caused by imaging geometry.

If the camera is tilted relative to a flat object, parallel edges can appear to converge even with a theoretically distortion-free lens.

Similarly, if a three-dimensional object changes height relative to a conventional perspective lens, its apparent size and position can change.

Software distortion correction does not automatically remove perspective error.

The camera should therefore be aligned mechanically before calibration. Otherwise, the calibration model may be forced to compensate simultaneously for lens distortion, camera tilt and object-plane error.

Distortion Is Not the Same as Chromatic or Spectral Focus Shift

A SWIR lens works across a broad wavelength range, so another optical issue is how focus behaves as wavelength changes.

If different spectral bands focus at slightly different axial positions, that produces spectral sharpness variation, not geometric barrel or pincushion distortion.

The two problems can interact because blurred edges are harder to localize accurately, but they should be diagnosed separately.

A machine measuring dimensions at multiple SWIR wavelengths should therefore test both geometric mapping stability and edge sharpness at the wavelengths actually used.

Why Edge Measurement Can Be More Difficult Than Center Measurement

Suppose the same 20 mm calibration feature is positioned at the image center and then near a corner.

If the center measurement is correct but the corner measurement changes consistently, lens distortion or calibration residual may be involved.

If both measurements fluctuate randomly, the problem could instead be poor edge contrast, noise, vibration or inadequate resolution.

This simple center-versus-edge test is extremely useful during SWIR system commissioning.

It separates systematic geometry error from general repeatability error.

Calibration Converts Pixel Coordinates Into Real-World Coordinates

Machine vision software often uses a calibration target of known geometry to determine the mapping between image coordinates and physical coordinates.

A simple system may use a scale such as:

0.10 mm/pixel

But one constant scale is accurate only if magnification is sufficiently uniform throughout the field.

A more advanced calibration estimates a spatial mapping that corrects radial and tangential distortion and compensates for camera pose relative to the measurement plane.

After calibration, the important metric becomes residual error: how far corrected measured points still deviate from their known physical positions.

For a serious dimensional SWIR system, residual calibration error is more useful than the raw distortion number alone.

A Calibration Model Cannot Recover Missing Optical Information

Software calibration is powerful but not unlimited.

If edge blur is severe, calibration cannot identify the true edge accurately.

If an object is represented by too few pixels, geometric correction cannot create missing spatial information.

If the lens produces large field-dependent blur or the optical assembly moves mechanically, a previously generated calibration may become invalid.

This is why the best engineering approach combines reasonably low optical distortion with calibration, rather than using software to compensate for an unsuitable optical system.

Distortion Calibration Must Use the Same Working Distance as Production

A lens calibration performed at one object distance should not automatically be assumed valid at another.

Focus position and lens internal geometry can change slightly with working distance, especially in lenses using manual focusing mechanisms.

If a machine is calibrated at 500 mm but operates at 300 mm, the geometric mapping should be verified again.

For production systems, calibrate at the actual working distance and focus setting that will be locked on the machine.

If the process deliberately uses several working distances, each should be validated separately.

Focus Adjustment Can Change Calibration

Mechanical focusing moves optical groups or changes lens-to-sensor geometry. Even small changes can affect magnification and potentially the distortion mapping.

This means a technician who refocuses the SWIR lens during maintenance can change the pixel-to-world relationship.

For geometry-sensitive equipment, focus should therefore be mechanically secured after calibration.

If focus is adjusted, the machine should either be recalibrated or checked against a traceable reference target before returning to production.

This is a crucial OEM maintenance requirement that is often overlooked.

Sensor Position and Lens Mount Repeatability Matter

The Kyptec Automation® SWIR portfolio uses C-Mount, providing a standardized mechanical interface across the range. But standardized thread geometry alone does not guarantee that two camera assemblies have identical optical alignment.

Sensor position, mount tolerance, lens seating, camera tilt and bracket repeatability can all influence calibration.

If an OEM intends to replace a lens or camera without recalibrating, that interchangeability should be validated experimentally.

For high-accuracy machines, a safer service policy is to require a calibration check whenever either optical component is replaced.

Distortion Matters to Region-of-Interest Placement

Even systems that do not perform dimensional measurement can be affected by geometric distortion.

Suppose the software defines a rectangular region of interest where a SWIR material response should occur. If the physical process feature is straight but the lens bends its image near the edge, a rigid rectangular ROI may include the wrong pixels.

This can create false rejects or missed defects.

Calibration can map a physical region into the appropriate distorted image coordinates, improving the relationship between process geometry and image analysis.

This is particularly useful for long edges, packaging compartments, PCB zones and structured arrays of objects.

Edge Geometry Matters for Defect Area Measurement

Some systems classify product quality partly by defect area.

If a contamination region covers 12 mm², it may pass; if it exceeds 15 mm², it may fail.

To compute physical area, the software needs reliable mapping between pixels and object space.

Distortion causes the physical area represented by a pixel to vary across the field unless the image is geometrically corrected.

This means an identical 20 × 20-pixel region may not correspond to exactly the same physical area at the image center and edge.

For area-based inspection, geometric calibration should therefore precede physical-area calculations.

Calibration Targets Should Cover the Full Usable Image

A small target placed only in the center cannot characterize edge distortion adequately.

Use a calibration pattern that covers the same region of the sensor used by production objects.

The target should have accurately known feature spacing and should be mounted in the same object plane as the inspected product.

For SWIR-only systems, target contrast must also remain strong at the actual calibration wavelength. A visually excellent printed pattern is not necessarily an ideal SWIR calibration target if its materials have poor infrared contrast.

The calibration target itself is part of the optical measurement system.

Reprojection Error Is a Useful Calibration Metric

After calibration, known target points can be projected through the model and compared with their detected image locations.

The difference is commonly summarized as reprojection error.

A small average error is useful, but OEMs should also examine maximum error and its spatial distribution.

If the average is 0.1 pixel but several edge points show 0.8 pixel residuals, those edge errors may still be unacceptable for a tight-tolerance machine.

The calibration report should therefore include center, mid-field and edge performance, not only one global average.

Convert Calibration Residual Into Object-Space Error

Pixel error becomes meaningful only when translated into millimetres or micrometres.

If the calibrated object sampling is 0.08 mm/pixel and maximum residual mapping error is 0.4 pixel:

0.08 × 0.4 = 0.032 mm

or approximately 32 µm.

The engineer can then compare 32 µm against the measurement tolerance.

This is the correct way to connect optical calibration with the machine's acceptance specification.

Measurement Uncertainty Is Larger Than Distortion Alone

An inspection system's total dimensional uncertainty can include:

lens distortion residual;

pixel localization uncertainty;

object sampling;

camera noise;

edge contrast;

mechanical vibration;

working-distance variation;

camera-to-object angle;

temperature;

and calibration-target uncertainty.

These contributions should be combined as an error budget rather than attributing every discrepancy to the lens.

A low-distortion SWIR camera lens reduces one important systematic error source, but total machine accuracy remains a system-level result.

Distortion Can Affect Robot or Ejector Coordinates

In a sorting or robotic inspection system, the camera may identify a defect and convert its image position into a physical target coordinate.

If geometric distortion is not corrected, coordinate error usually changes with field position.

A target near the image center may be picked or rejected correctly while the same target near the edge is missed.

This pattern is a classic sign that the camera-to-machine coordinate mapping requires better calibration.

For SWIR systems performing material classification followed by physical removal, distortion therefore becomes relevant even when the classifier itself is unaffected.

Low-Distortion Optics Reduce Calibration Burden

The purpose of choosing a lower-distortion lens is not necessarily to eliminate calibration.

It is to begin with a more geometrically faithful image so the calibration model needs to correct less.

Smaller correction generally helps preserve useful image area, reduces interpolation displacement and can improve robustness when the machine is replicated.

This is where the lower published TV distortion of models such as the Kyptec Automation® KL-1414 and Kyptec Automation® KL-1416 can be especially valuable for geometry-sensitive SWIR systems.

Why Kyptec Automation® Is a Strong SWIR Platform for Geometry-Sensitive OEM Systems

The Kyptec Automation® SWIR Camera Lens collection provides more than a sequence of focal lengths. It gives OEMs a practical way to balance coverage, working distance and geometric performance across a consistent 900–1700 nm lens family. The current range covers 8.5 mm, 12.5 mm, 25 mm, 35 mm and 50 mm with F1.4, 2 MP, 2/3-inch format and C-Mount.

The published distortion progression is particularly useful for engineering selection: the wider Kyptec Automation® KL-1408 is specified at approximately -1.1% TV distortion, while the 12.5 mm and 25 mm models are around -0.23% and -0.25%, respectively. The Kyptec Automation® KL-1414 35 MM SWIR Camera Lens is specified around -0.05%, and the Kyptec Automation® KL-1416 50 MM SWIR Camera Lens around -0.08%. These values allow buyers to evaluate distortion alongside FOV rather than treating focal length as the only optical selection criterion.

For a classification machine, the wider model may be entirely appropriate. For calibrated dimensional or coordinate inspection, the lower-distortion longer focal lengths may provide a stronger starting point when the required FOV and machine envelope allow them. This is precisely why a multi-focal-length specialized SWIR portfolio is valuable to OEM buyers.

Frequently Asked Questions About SWIR Camera Lens Distortion

1. What does TV distortion mean on a SWIR camera lens datasheet?

TV distortion expresses how much image geometry differs from ideal geometric projection. It primarily describes changes in magnification across the image that cause straight physical lines to appear curved. It should not be confused with image sharpness, resolution or focus quality.

2. Is lower TV distortion always better?

For geometric measurement, coordinate extraction and calibration-sensitive inspection, lower distortion is generally advantageous. However, it should not override essential requirements such as FOV and working distance. A lens with extremely low distortion is not useful if it cannot cover the required object area.

3. Does 1% lens distortion mean every measurement is wrong by exactly 1%?

No. Lens distortion is spatially varying and generally increases away from the optical center. A datasheet percentage does not mean every dimension measured anywhere in the image has exactly that error. Production accuracy should be established through full-field calibration and residual-error testing.

4. Does lens distortion affect SWIR material classification?

Usually much less than it affects dimensional measurement. If the classifier analyzes sufficiently large material regions, moderate distortion may have little impact. It becomes important when the distortion changes ROI placement, mixed pixels or the physical coordinates associated with a classified object.

5. Why are measurements often less accurate near the edge of a SWIR image?

Radial distortion tends to increase toward the image edge, and lenses can also show reduced edge sharpness compared with the center. A calibration model can correct systematic geometric deformation, but sufficient local contrast and resolution are still required for accurate edge localization.

6. Can software completely remove SWIR lens distortion?

Software can correct a stable, well-characterized distortion model very effectively, but it cannot recover missing optical resolution, correct unstable mechanical alignment or fully compensate for imaging conditions that change after calibration. Lower-distortion optics plus calibration is generally stronger than depending on correction alone.

7. What is the difference between lens distortion and camera perspective error?

Lens distortion originates from the optical design and changes magnification across the field. Perspective error originates from camera-object geometry, such as a tilted camera or changing object depth. A calibration model may account for both, but they are physically different error sources.

8. Does changing working distance affect distortion calibration?

It can affect the geometric mapping and image magnification, particularly if focus must also be adjusted. A calibration generated at one working distance should therefore be verified before being applied at another. Measurement systems should ideally operate at a fixed, mechanically controlled working distance.

9. Does refocusing a SWIR camera lens require recalibration?

For geometry-sensitive inspection, it is good practice to check or repeat calibration after significant focus adjustment. Refocusing can slightly change magnification and the relationship between sensor pixels and object coordinates. Focus should normally be locked after the production calibration is established.

10. Which Kyptec Automation® SWIR lens has the lowest published TV distortion?

Within the current five-lens portfolio, the Kyptec Automation® KL-1414 35 MM SWIR Camera Lens has a published TV distortion of approximately -0.05%, making it particularly interesting for low-distortion SWIR applications where its narrower FOV is compatible with the machine geometry.

11. What is the published TV distortion of the Kyptec Automation® KL-1408 8.5 MM SWIR Camera Lens?

The official datasheet specifies approximately -1.1% TV distortion for the Kyptec Automation® KL-1408 8.5 MM SWIR Camera Lens. This should be evaluated together with its major advantage: the broad field of view offered by the shortest focal length in the range.

12. Is the 12.5 mm SWIR lens better than 8.5 mm for measurement?

It can be a stronger choice when the required FOV permits it because the Kyptec Automation® KL-1410 12.5 MM SWIR Camera Lens has much lower published distortion than the 8.5 mm model while also providing stronger object sampling for the same sensor. However, actual measurement accuracy must still be validated after calibration.

13. Why might an OEM choose the 25 mm SWIR lens for calibrated inspection?

The Kyptec Automation® KL-1412 25 MM SWIR Camera Lens can be useful where the inspection region is smaller and greater sensor utilization is desirable. Its approximately -0.25% published TV distortion provides a useful low-distortion geometry for controlled object inspection when the required working distance and FOV match the machine.

14. Is the 50 mm SWIR lens suitable for precision edge geometry?

The Kyptec Automation® KL-1416 50 MM SWIR Camera Lens has approximately -0.08% published TV distortion and can be a strong candidate for tighter geometry-sensitive inspection. Its longer focal length requires sufficient working distance and provides a narrower field, so the machine must allow adequate positional margin.

15. How should I test distortion before approving a SWIR lens for production?

Image an accurately manufactured calibration target at the real working distance using the final camera, lens, aperture and focus setting. Measure deviations at the image center, mid-field and corners. Generate the intended calibration model and then measure residual error using independent target positions that were not used simply to fit the model.

16. How many calibration points should a SWIR measurement system use?

There is no universal number because it depends on field size, distortion complexity and required accuracy. The important requirement is that calibration points adequately cover the full production region rather than being concentrated near the center. Higher-accuracy applications generally benefit from dense, well-distributed reference points.

17. Can distortion change when I replace the lens with another unit of the same model?

Small unit-to-unit optical and mechanical differences are possible in any production optical system. For ordinary detection this may be insignificant, but a high-accuracy calibrated machine should verify calibration after lens replacement rather than assuming two physical units are geometrically identical.

18. Why does my SWIR measurement machine pass calibration in the center but fail at the corners?

This usually indicates that a single global scale factor is not adequate, the lens distortion model is incomplete, the calibration target does not cover the full field, or edge localization quality is worse than expected. A full-field geometric calibration should be performed and the maximum residual error evaluated separately from the average error.

19. What specifications should an OEM provide when selecting a low-distortion SWIR camera lens?

Provide sensor dimensions, required horizontal and vertical FOV, working distance, smallest measurable feature, allowable dimensional or coordinate error, expected object-position range, operating wavelength, whether full-field measurement is required and whether software calibration will be used. These inputs make it possible to balance focal length and distortion against the real measurement tolerance.

20. Why is Kyptec Automation® a strong choice for low-distortion SWIR inspection?

Kyptec Automation® offers a dedicated SWIR Camera Lens portfolio with multiple focal lengths and published distortion characteristics rather than one generic SWIR geometry. The range extends from wide 8.5 mm coverage to lower-distortion 35 mm and 50 mm configurations while maintaining a common 900–1700 nm, 2 MP, 2/3-inch, F1.4 and C-Mount platform. This gives OEMs a practical way to select the optical geometry according to both FOV and calibration accuracy rather than compromising one requirement blindly.

Conclusion

SWIR camera lens distortion becomes important whenever an inspection system converts an infrared image into physical geometry. Material classification can often tolerate a modest amount of distortion, but dimensional measurement, coordinate extraction, calibrated defect mapping, robot guidance, region-of-interest placement and physical-area calculation require a much stronger understanding of how the lens maps the object plane onto the sensor. Kyptec Automation® already notes in its broader machine-vision guidance that barrel or pincushion distortion can influence dimensional accuracy and may require calibration or lower-distortion optics for precision measurement.

TV distortion should therefore never be treated as an isolated datasheet number. The OEM should convert it into the context of actual field width and tolerance, test geometric behavior from the image center to the corners, calibrate at the real working distance, lock focus and mounting after calibration, and calculate residual error in object-space units. A machine requiring ±0.05 mm dimensional accuracy needs a fundamentally different distortion strategy from one whose only purpose is to classify a 50 mm material sample.

The Kyptec Automation® SWIR Camera Lens collection gives buyers useful flexibility for making that trade-off. The wide Kyptec Automation® KL-1408 8.5 MM SWIR Camera Lens provides substantial coverage and is officially specified at approximately -1.1% TV distortion. The 12.5 mm and 25 mm models provide substantially lower published distortion while narrowing the field, and the Kyptec Automation® KL-1414 35 MM SWIR Camera Lens and Kyptec Automation® KL-1416 50 MM SWIR Camera Lens offer particularly low published distortion for tighter geometry-sensitive applications. This progression lets an OEM select according to the actual balance between coverage, stand-off, sensor utilization and geometric accuracy.

For machine builders, the strongest design principle is to specify allowable object-space error first and choose the SWIR lens second. Define how accurately the machine must locate or measure a feature, identify where that feature can appear across the sensor, calculate the necessary FOV, then compare the distortion behavior of the focal lengths that satisfy the mechanical geometry. After selection, validate the complete camera-lens assembly with a full-field calibration target and use the remaining residual error—not appearance alone—as the final acceptance criterion. When this process is followed, Kyptec Automation® SWIR Camera Lenses provide a strong and technically coherent foundation for 900–1700 nm inspection systems where material visibility and geometric accuracy must work together.