SWIR Lens MTF Explained: How Resolution, Pixel Pitch and Spatial Frequency Determine Real Inspection Detail

When an industrial buyer asks whether a SWIR camera lens is “sharp enough,” the question is more complicated than it first appears. A lens does not either resolve an object perfectly or fail completely. As features become progressively smaller, the optical system transfers less contrast from the real object to the sensor. Fine dark and bright structures that are strongly separated on the object gradually appear more similar in the captured image. Modulation Transfer Function, or MTF, describes this loss of contrast as spatial detail becomes finer, making it one of the most useful concepts for understanding whether a SWIR imaging system can actually distinguish the defects, edges, particles, coatings or material boundaries required by an automated inspection.

For engineers searching for a SWIR lens MTF, high resolution SWIR camera lens, SWIR lens resolution, SWIR lens for small defect detection, 2 MP SWIR lens, 900–1700 nm SWIR lens, or C-Mount SWIR camera lens, the practical objective is not to chase the highest theoretical resolution number. The goal is to match optical performance, sensor pixel pitch, field coverage and the smallest required inspection feature so sufficient contrast reaches the image-processing algorithm. The Kyptec Automation® SWIR Camera Lens collection provides 8.5 mm, 12.5 mm, 25 mm, 35 mm and 50 mm focal-length choices within a dedicated 900–1700 nm, 2 MP, 2/3-inch, F1.4 and C-Mount platform, allowing optical geometry to be selected while maintaining a consistent SWIR lens family.

What MTF Means in a SWIR Camera Lens

MTF measures how effectively an optical system transfers contrast from an object to its image at different spatial frequencies. A large feature with broad black and white areas represents relatively low spatial frequency. Very fine alternating lines represent higher spatial frequency. An ideal optical system would reproduce both with the same contrast, but real lenses increasingly reduce contrast as structures become finer.

If an object contains perfectly black and perfectly white bars, the object contrast is high. After those bars pass through the lens, the darkest image region may become lighter and the brightest region darker. The difference between them decreases. At sufficiently fine spacing, the two may become almost indistinguishable even though technically some information remains.

This distinction is fundamental to machine vision because an inspection algorithm rarely benefits from detail that exists only mathematically. It needs enough contrast to detect an edge or classify a feature reliably in the presence of sensor noise, illumination variation, object movement and production tolerance.

MTF therefore provides a more useful way to think about SWIR lens resolution than simply asking whether the lens can “see” a particular line pair.

Spatial Frequency: The Language Used to Describe Fine Optical Detail

Spatial frequency expresses how rapidly brightness changes across an image. In lens engineering it is commonly described in line pairs per millimetre (lp/mm) at the sensor plane. One line pair consists of one dark line and one bright line.

Low spatial frequency represents large structures. High spatial frequency represents finer structures packed closely together.

A lens may transfer very high contrast at 10 lp/mm, somewhat lower contrast at 30 lp/mm and substantially lower contrast at 60 lp/mm. This does not mean the lens suddenly stops resolving at one frequency. Optical resolution normally decreases gradually.

That is why a meaningful SWIR lens evaluation should ask two questions together: at what spatial frequency is performance being discussed, and how much contrast remains at that frequency?

A high lp/mm claim without contrast information provides an incomplete picture of useful inspection performance.

Why “2 Megapixel Lens” and “2 Megapixel Camera” Do Not Mean the Same Thing

Megapixels describe the number of sensor samples or the resolution class for which an optical product is intended; MTF describes how contrast is transferred across different levels of detail. These are related but fundamentally different concepts.

A camera with two million pixels records approximately two million discrete sampling locations. The lens creates the analogue optical image that those pixels sample. If the lens substantially attenuates fine detail before it reaches the sensor, additional digital sampling cannot restore the lost contrast.

Conversely, strong optical performance does not automatically create inspection detail if the sensor has insufficient sampling or if the required field of view spreads its pixels across too much object area.

The current Kyptec Automation® SWIR Camera Lens portfolio is specified at 2 MP resolution, but buyers should interpret that specification together with sensor format, field size, target detail and real application testing rather than treating “2 MP” as an isolated guarantee of smallest detectable defect.

Pixel Pitch Creates a Sampling Limit, Not an Optical Quality Guarantee

Pixel pitch is the physical spacing between adjacent sensor pixels. Smaller pixels allow the detector to sample the optical image more densely. The theoretical sensor sampling limit can be estimated from pixel pitch using the Nyquist relationship:

Nyquist frequency ≈ 1 ÷ (2 × pixel pitch)

When pixel pitch is expressed in millimetres, the result is cycles or line pairs per millimetre.

For example, a hypothetical sensor with a 5 µm pixel pitch has a pitch of 0.005 mm. Its approximate Nyquist frequency is:

1 ÷ (2 × 0.005) = 100 lp/mm

This does not mean the complete system delivers perfect 100 lp/mm inspection performance. It only indicates the theoretical digital sampling boundary. The lens must still transfer useful contrast at relevant frequencies, and the application needs enough pixels across the target feature.

The practical system resolution is therefore controlled by the interaction of lens MTF + sensor sampling + object magnification + illumination + motion + focus.

Why Designing Exactly at Nyquist Is Usually Too Aggressive for Production Inspection

Nyquist is a mathematical sampling boundary, not a recommended industrial operating target. If a production defect is represented extremely close to the sampling limit, its apparent shape and contrast can change dramatically depending on its alignment with the pixel grid. Small focus errors, vibration or motion blur can then cause detection to become unstable.

A stronger design leaves margin.

Instead of asking whether the sensor theoretically samples the smallest feature, engineers should determine whether the feature occupies enough pixels and retains sufficient contrast after the complete imaging chain has acted on it. This is especially important in SWIR material inspection, where classification may require both spatial information and reliable intensity information inside the feature.

An inclusion occupying many well-defined pixels is much easier to classify than one represented by a weak two-pixel transition, even if both technically satisfy a theoretical sampling calculation.

Object-Space Resolution and Image-Space Resolution Must Not Be Confused

Lens MTF is frequently expressed at the image plane in lp/mm, while an inspection specification is usually defined at the object: a 0.3 mm crack, a 0.5 mm contaminant, a 1 mm coating defect or another physical feature size.

The relationship between these two spaces depends on magnification.

A feature that is 1 mm wide on the object does not necessarily produce a 1 mm feature on the sensor. Its image size depends on optical magnification. This means a lens may support a certain image-plane spatial frequency, yet whether that capability resolves a particular production defect depends on the geometry of the complete system.

This is one reason why buying a lens only from an lp/mm specification can be misleading. The engineer must translate the actual object feature into the corresponding sensor-plane feature size before determining what optical frequency matters.

Contrast at the Required Feature Size Is More Important Than Maximum Resolution

Imagine that a system must detect fine contamination. It is tempting to select optics according to the smallest possible theoretical line spacing. In practice, the more useful question is whether the target produces enough image modulation to separate it reliably from acceptable material.

Suppose one optical configuration technically resolves an extremely fine feature but reproduces it at very low contrast. Another configuration does not reach the same extreme limiting frequency but produces considerably stronger contrast around the defect sizes that actually matter. The second configuration may perform better in production.

Machine vision therefore benefits from usable contrast, not record-setting limiting resolution.

For SWIR applications, this is especially important because spectral contrast itself may already be relatively subtle. Allowing spatial contrast to deteriorate unnecessarily can reduce the separation between material classes or defects.

MTF at the Center and Edge of the 2/3-Inch Image Can Differ

Lens performance is not always identical across the complete image field. The optical axis generally represents the least demanding part of the image. Toward the edges, off-axis aberrations and geometry can make maintaining fine-detail contrast more challenging.

For a 2/3-inch SWIR camera lens, center performance alone is therefore insufficient if products can appear anywhere across the sensor.

A defect-detection machine should test the same fine feature at multiple image positions. If a defect is detected consistently in the center but becomes weak near a corner, the true usable inspection field is smaller than the nominal sensor coverage.

This type of testing is particularly important for wide-field imaging, where edge regions form a substantial part of the production inspection zone.

Why Shorter Focal Lengths Can Make Resolution Demands More Difficult

Focal length itself does not determine MTF performance in isolation, but shorter focal lengths are often selected to capture wider fields from a given working distance. Once the field becomes wider, each sensor pixel represents a larger area on the object.

That geometric change can create a stricter defect-detection problem even when the camera and lens resolution remain unchanged.

The Kyptec Automation® KL-1408 8.5 MM SWIR Camera Lens is the widest focal-length option in the current portfolio. In applications where broad conveyor or object coverage is necessary, its geometry can be valuable, but system designers should verify that the smallest required feature still occupies sufficient pixels and retains enough optical contrast across the field.

The correct question is therefore not whether 8.5 mm is “less sharp” than a longer focal length. The question is whether the wider field created by the selected geometry provides adequate object-space resolution for the inspection.

Why Image Scale Can Be More Important Than Adding Camera Pixels

Suppose a target occupies only a small percentage of the image. Increasing camera pixel count may improve sampling, but changing optical geometry so that the relevant target fills more of the available sensor can sometimes provide a more efficient improvement.

A longer focal length or altered working geometry can increase image scale, allocating more pixels to the same object. This reduces the required object-space detail represented by each pixel and can move the important inspection feature into a more comfortable portion of the lens-sensor resolution budget.

For a controlled inspection region, the Kyptec Automation® KL-1412 25 MM SWIR Camera Lens can be evaluated where a medium field provides a more efficient use of the available 2/3-inch sensor than an unnecessarily wide view.

This is why resolution problems should not automatically be solved by buying a higher-pixel camera. Sometimes the more fundamental issue is that the target has been imaged too small.

How MTF Relates to Edge Detection

Many automated inspection algorithms begin with edges. Scratches, package boundaries, cracks, component outlines and coating transitions all contain spatial changes in brightness.

When a lens loses high-frequency contrast, these transitions become less steep. A sharp black-to-white edge turns into a gradual grey transition across multiple pixels. The algorithm may still locate the edge, but position accuracy and defect contrast can deteriorate.

This has direct consequences for measurements. If an inspection system determines width, spacing, alignment or edge location, reduced MTF can increase uncertainty even when objects remain visibly recognizable.

For SWIR inspection, edge quality can become even more important when material contrast is produced spectrally rather than through strong visible-light texture. The optical system needs to preserve both the material intensity difference and the spatial boundary separating regions.

MTF50, MTF10 and Limiting Resolution Are Not Interchangeable

When optical performance is described using MTF, the percentage attached to the value matters. MTF50 refers to the spatial frequency where contrast has fallen to 50% of its low-frequency reference. MTF10 corresponds to a much lower remaining contrast. Limiting resolution may be defined differently again.

A lens may therefore have a high numerical resolution at low contrast while delivering much lower frequency at strong contrast.

For industrial buyers, this means two specifications expressed simply as “X lp/mm” may not be comparable unless the associated MTF criterion is known.

The application should define the contrast level required for reliable detection rather than treating the smallest barely visible structure as the relevant system specification.

Why SWIR Wavelength Can Influence Fine-Detail Performance

Diffraction and optical aberrations are wavelength dependent. SWIR wavelengths are longer than visible wavelengths, so assumptions based purely on visible-light image sharpness should not automatically be transferred to a 900–1700 nm system.

The inspection should be validated at the wavelength or wavelengths actually used in production. A lens-camera combination may show different fine-detail contrast when illumination changes substantially across the SWIR band.

This does not mean engineers need a laboratory MTF bench for every production machine. A practical approach is to image representative fine structures or calibration targets under each actual SWIR illumination condition and confirm that the target defects retain sufficient contrast.

This is different from simply verifying that the lens transmits the selected wavelength. Transmission determines how much signal reaches the sensor; MTF describes how well spatial contrast is preserved. Both matter, but they answer different engineering questions.

Aperture Can Change the Resolution Trade-Off

Changing aperture influences several imaging properties simultaneously. Opening the aperture increases light collection but may make the system more sensitive to certain aberrations and reduce depth of field. Closing the aperture can improve some aberration-related behaviour and increase depth of field, but eventually diffraction becomes increasingly important while signal decreases.

There is therefore no universal aperture setting that produces maximum inspection performance.

The Kyptec Automation® SWIR Camera Lens family is specified at F1.4, providing substantial available light collection for compatible systems. During integration, the operating aperture should be validated according to actual target depth, illumination level, exposure time and required fine-detail contrast.

For an MTF-sensitive inspection, aperture should be selected by measuring the resulting image performance rather than simply operating permanently at the widest available setting.

Motion Blur Acts Like Another MTF Loss

Even an excellent lens cannot preserve fine inspection detail if the target moves significantly during exposure. Motion blur spreads each object feature over additional pixels and reduces contrast in the direction of movement.

This can produce a misleading diagnostic situation. Engineers may believe the lens lacks resolution when the real limitation is exposure time.

For high-speed SWIR inspection, evaluate the static image first and then repeat at production speed. If fine details disappear only during movement, shorter exposure, stronger illumination or changes to transport speed may be required.

Optical MTF, sensor sampling and motion performance should therefore be considered as a combined system MTF rather than isolated specifications.

Defocus Can Destroy High-Frequency Contrast Before the Image Looks Obviously Blurred

One important characteristic of MTF is that fine features usually deteriorate before large structures do. A slight focus error may leave an image looking acceptable to the human eye while significantly reducing contrast at high spatial frequencies.

This is why manual visual focusing is not always sufficient for a precision SWIR inspection machine.

A better approach is to optimize focus using the actual smallest production feature or a suitable high-frequency target under the real SWIR illumination. Once optimum focus is found, the mechanical system should hold that setting reliably.

If object height varies, test the complete depth range. The lens may provide excellent fine-detail contrast at nominal focus but insufficient performance at the maximum permitted displacement.

Using Longer Focal Lengths to Allocate More Sensor Sampling to a Small Target

Where the inspection region is relatively narrow, a longer focal length can help dedicate more of the sensor to the target.

The Kyptec Automation® KL-1414 35 MM SWIR Camera Lens provides a tighter field option in the current portfolio. In applications where the installation geometry permits greater stand-off, this type of focal length can be evaluated for localized material zones, small components or defect regions that would occupy too few pixels in a wider image.

The Kyptec Automation® KL-1416 50 MM SWIR Camera Lens extends that geometry further for comparatively narrow inspection fields.

Neither focal length automatically increases intrinsic system resolution. Their advantage is that, under suitable geometry, the object can occupy more of the sensor. That improved image scale can make the available optical and pixel resolution more useful to the application.

How to Build a Practical Resolution Budget for a SWIR Inspection

A good resolution budget begins with the smallest real-world feature the machine must detect reliably. Next determine the required field of view and available camera pixel count. This provides the nominal object-space sampling.

Then add margin. Do not design so the smallest defect occupies the theoretical minimum number of pixels. Account for optical MTF, edge position, sensor noise, manufacturing variation, motion blur, imperfect focus and field-dependent performance.

After that, validate the lens at the actual SWIR wavelength, aperture and image position. Test the defect at the center and edges, at minimum and maximum expected object distance, and at full production speed.

The final criterion should be repeatable detection probability, not whether a calibration target can barely be resolved under ideal conditions.

Why Kyptec Automation® Is a Practical SWIR Lens Platform for Resolution-Conscious System Design

Kyptec Automation® provides a focused SWIR Camera Lens family rather than forcing engineers to solve every imaging geometry with one focal length. The current portfolio covers 8.5 mm, 12.5 mm, 25 mm, 35 mm and 50 mm, while maintaining a published 900–1700 nm wavelength range, 2 MP resolution class, 2/3-inch sensor format, F1.4 aperture and C-Mount architecture.

For an OEM or system integrator, this gives useful design flexibility. A wide inspection station can evaluate the shorter focal lengths, while a smaller target region can use a tighter field so more available sensor samples are devoted to meaningful detail. Because the models belong to the same dedicated Kyptec Automation® SWIR Camera Lens category, focal-length selection can remain directly tied to the application's resolution budget and machine geometry.

The value is not simply having multiple focal lengths. It is being able to choose the geometry that helps the complete lens-sensor system preserve the level of detail the production decision actually requires.

Frequently Asked Questions About SWIR Lens MTF, Resolution and Spatial Frequency

1. What is MTF in a SWIR camera lens?

MTF, or Modulation Transfer Function, describes how much image contrast a lens preserves as spatial detail becomes finer. Low-frequency features such as broad edges normally retain stronger contrast, while very fine line patterns lose progressively more contrast. For SWIR machine vision, MTF helps determine whether small defects or boundaries will remain strong enough for reliable automated detection rather than merely being theoretically visible.

2. Is higher lp/mm always better for a SWIR lens?

Not by itself. A high line-pairs-per-millimetre number is useful only when the associated contrast level and test conditions are known. A lens may technically resolve a very fine pattern at extremely low contrast while producing stronger, more useful contrast at a lower frequency. Industrial inspection should prioritize the spatial frequencies corresponding to the actual defect size and the contrast needed by the algorithm.

3. What is the difference between MTF and lens resolution?

Resolution often describes the ability to distinguish fine detail, while MTF describes how contrast changes continuously as detail becomes finer. MTF provides more information because it shows that optical performance does not abruptly stop at one resolution value. Two lenses with similar limiting resolution can behave differently at the moderate spatial frequencies that matter most to a production inspection.

4. What spatial frequency should a SWIR machine-vision lens support?

The relevant spatial frequency depends on sensor pixel pitch, magnification and the smallest object feature that must be detected. There is no universal lp/mm requirement for every SWIR system. The engineer should translate the real defect size into its sensor-plane image size and determine how much optical contrast is required at that frequency, then validate the complete camera-lens system under actual operating conditions.

5. How does MTF affect small defect detection?

Small defects contain higher spatial-frequency information than large objects. If lens MTF is low at those frequencies, a fine scratch or small contaminant loses contrast and can become difficult to separate from its surroundings. Detection may still work under ideal conditions but become unstable when noise, motion, focus variation or product differences are introduced. Strong usable MTF around the required defect scale therefore provides valuable inspection margin.

6. What does MTF50 mean for a camera lens?

MTF50 is the spatial frequency at which the system retains approximately half of the reference contrast. It is often useful as an indicator of perceived or practical sharpness, but it should not be treated as the only performance measure. Industrial machine vision may care about a different contrast threshold depending on how much modulation an algorithm needs to distinguish the feature reliably.

7. Is MTF10 the same as maximum usable resolution?

Not necessarily. MTF10 indicates a point where only about 10% of the reference contrast remains, which can correspond to relatively fine detail but weak modulation. Whether that information is usable depends on sensor noise, illumination stability, algorithm sensitivity and production variation. An industrial system should normally avoid relying on barely transferred detail if reliable defect detection is required.

8. How does diffraction affect SWIR lens resolution?

Diffraction places a physical limit on optical resolution and becomes more significant as the aperture is stopped down. Because diffraction depends on wavelength, the longer wavelengths used in SWIR imaging can make the trade-off particularly relevant. Closing the aperture may increase depth of field and reduce some aberrations, but excessive stopping down can reduce fine-detail contrast. The optimum setting should therefore be verified at the actual operating wavelength.

9. Why does a sharp SWIR image sometimes still miss small defects?

Overall visual sharpness is dominated by larger and medium-size structures, while a small defect may depend on considerably higher spatial frequencies. An image can appear clear to an operator yet transfer insufficient contrast at the exact feature scale required by the inspection. Motion blur, imperfect focus or inadequate object sampling can worsen this effect. Resolution should therefore be validated using the smallest real target, not general visual appearance.

10. How can I test SWIR lens resolution without a laboratory MTF bench?

Use a suitable resolution target or, preferably, representative production features under the actual SWIR illumination, camera, aperture and working geometry. Compare fine-detail contrast at the center and edges of the field, then repeat at the expected object-distance range and production speed. Laboratory MTF measurements provide deeper characterization, but application-level testing can determine whether the system has enough practical resolution for the intended inspection.

11. Why should MTF be checked at multiple positions across the sensor?

A lens can transfer different levels of contrast at the center and edge of the image. If products or defects can appear anywhere within a 2/3-inch inspection field, center-only measurements can overstate real production capability. Full-field validation determines whether the complete sensor area provides sufficiently consistent fine-detail performance for reliable classification or dimensional analysis.

12. Does a longer focal length automatically give higher MTF?

No. Focal length and MTF describe different aspects of the optical system. A longer focal length may create greater target magnification or a narrower field at a particular working distance, allowing the object to occupy more sensor pixels. That can improve practical defect representation even without changing intrinsic lens MTF. Models such as the Kyptec Automation® KL-1414 35 MM SWIR Camera Lens should therefore be selected for appropriate geometry rather than because longer focal length automatically means better optical resolution.

13. Can motion blur reduce MTF even when the lens is correctly focused?

Yes. Motion during exposure spreads image information along the direction of movement and reduces contrast at fine spatial frequencies. This acts similarly to another low-pass filter in the imaging chain. A static resolution test can therefore look excellent while the production image loses small defects at conveyor speed. High-speed SWIR inspection should always be validated under actual motion conditions.

14. What is system MTF and why is it more important than lens MTF alone?

System MTF describes the combined spatial-frequency response of the imaging chain rather than the lens in isolation. Lens optics, sensor sampling, pixel aperture, focus, motion, wavelength and image processing can all influence final contrast. A premium lens cannot compensate for severe motion blur or poor sampling, while a high-resolution sensor cannot restore contrast already lost optically. Production performance depends on the complete system.

15. How much resolution margin should I allow when selecting a SWIR camera lens?

There is no universal percentage because inspection difficulty varies, but the system should not be designed so the smallest critical feature sits at the theoretical optical or sampling limit. Provide enough pixel coverage and optical contrast to accommodate focus drift, positioning tolerance, sensor noise, illumination variation, motion and normal product variability. The safest purchasing method is to calculate a conservative resolution budget and then verify it with representative defects before freezing the OEM design.

Conclusion

Real SWIR inspection detail is determined by far more than megapixel count. The lens must transfer contrast at the spatial frequencies corresponding to the features that matter, the sensor must sample those features adequately, the image scale must allocate enough pixels to the target, and focus, wavelength, aperture and motion must preserve the available information through production conditions. MTF provides the framework that connects those variables.

For buyers, this means a high-resolution SWIR camera lens should be selected by asking how much useful contrast the complete imaging system provides at the required feature scale, not merely by comparing a maximum lp/mm figure or camera megapixel count. Low-frequency image quality can look excellent even while tiny defects are disappearing, so validation must be performed with the smallest relevant features and across the complete usable image field.

The Kyptec Automation® SWIR Camera Lens collection gives OEMs and machine-vision integrators a focused set of 8.5 mm, 12.5 mm, 25 mm, 35 mm and 50 mm options within a verified 900–1700 nm, 2 MP, 2/3-inch, F1.4, C-Mount platform. By selecting the focal length according to image scale and field requirements, then validating spatial contrast under actual SWIR illumination and production conditions, users can build inspection systems around the detail that matters in the real application rather than relying only on nominal resolution specifications.