SWIR Camera Lens MTF and Resolution Explained: How Optical Contrast Determines the Smallest Detectable Industrial Defect

When an industrial buyer asks how small a defect a SWIR imaging system can detect, camera resolution is often the first specification considered. A 2 MP camera may appear to offer a straightforward answer: calculate the number of pixels across the field of view, determine millimetres per pixel, and assume that any feature larger than a few pixels will be detectable. In practice, SWIR camera lens resolution is not determined by pixel count alone. A defect can occupy several sensor pixels and still disappear if the optical system transfers too little contrast at that feature size. This is where modulation transfer function, or MTF, becomes fundamental. MTF describes how effectively a lens preserves contrast as object details become progressively finer, making it one of the most useful concepts for understanding the relationship between lens resolution, sensor resolution, pixel pitch, field of view and the smallest industrial defect that can actually be inspected.

For 900–1700 nm imaging, this relationship becomes particularly important because the inspection target may already have lower optical contrast than a conventional visible feature. A moisture boundary, material inclusion, microcrack, thin contaminant, coating variation or small spectral anomaly may generate only a modest difference from its surroundings. If the lens reduces that difference further before the image reaches the sensor, the defect can become statistically indistinguishable from noise or normal product variation. The Kyptec Automation® SWIR Camera Lens collection is designed specifically for the 900–1700 nm range and currently includes 8.5 mm, 12.5 mm, 25 mm, 35 mm and 50 mm focal lengths with 2 MP resolution class, 2/3-inch format, F1.4 aperture and C-Mount. For an OEM, the important task is to connect those optical specifications to the actual contrast and spatial-frequency requirement of the inspection application.

What MTF Really Means in a SWIR Camera Lens

MTF describes how much contrast an optical system retains when imaging structures of different spatial sizes. Large features are relatively easy for a lens to reproduce because their transitions occur slowly across the image. As features become progressively smaller and more closely spaced, optical aberrations, diffraction, defocus and other effects reduce the contrast reaching the sensor. Eventually the object may still contain a physical difference, but the image no longer contains enough separation for reliable detection. This is why MTF should be thought of as contrast transfer versus spatial frequency, not simply as a theoretical optical specification.

Consider an object pattern containing equally bright and dark regions. Object contrast can be expressed as C = (Imax − Imin) / (Imax + Imin). If the original feature has 80% contrast but the lens produces only 40% contrast at the corresponding spatial frequency, the MTF at that frequency is approximately 0.5 because half of the original contrast remains. If the same lens transfers only 8% contrast at a much finer spatial frequency, those small features may become extremely difficult for the inspection algorithm even though the sensor technically contains enough pixels to sample them. In real industrial SWIR imaging, the target is rarely a perfect black-and-white test pattern, making adequate contrast transfer even more important.

Resolution and Detection Are Not the Same Specification

Optical resolution generally describes the ability to distinguish fine spatial detail, while defect detection describes whether the complete imaging system can make a sufficiently reliable production decision about a particular feature. These are related but not identical. A lens may resolve two fine structures under controlled conditions yet transfer insufficient contrast for a machine-learning or threshold-based algorithm to classify them reliably under production noise, material variation and motion. Conversely, a comparatively large defect with strong SWIR contrast may remain easy to detect even without exceptionally high spatial resolution.

For machine builders, the useful question is therefore not simply “What is the maximum resolution of this lens?” but “How much contrast remains at the spatial scale corresponding to my minimum defect, at my actual wavelength, aperture, working distance and field position?” This converts MTF from an optical laboratory concept into a practical machine-design parameter.

Why a 2 MP SWIR Lens Does Not Mean Every 2 MP Camera Will Produce the Same Inspection Result

A lens specified for 2 MP optical performance should be evaluated together with the sensor rather than treated as an independent guarantee. Two 2 MP SWIR cameras can have different sensor dimensions and therefore different pixel pitches. A smaller physical sensor containing the same number of pixels has finer pixels and samples higher spatial frequencies. The lens must transfer usable contrast at those frequencies if the additional sampling is to provide practical benefit.

The current Kyptec Automation® SWIR range is positioned around 2 MP resolution and 2/3-inch sensor format. For example, the Kyptec Automation® KL-1410 12.5 MM SWIR Camera Lens is specified for 900–1700 nm, 2 MP, F1.4, 2/3-inch format and C-Mount. This is a strong starting architecture for matching an industrial SWIR camera, but minimum detectable defect should still be calculated from the actual sensor, FOV and required contrast margin.

Pixel Pitch Determines the Sensor's Sampling Frequency

Pixel pitch describes the physical spacing between neighbouring sensor pixels. If pixel pitch is (p), a simplified Nyquist spatial frequency is fN = 1/(2p), with pixel pitch expressed in millimetres when frequency is required in line pairs or cycles per millimetre. For a hypothetical 10 µm pixel, (p = 0.010) mm, giving a Nyquist frequency of approximately 50 cycles/mm. A 5 µm pixel would theoretically sample up to approximately 100 cycles/mm. The smaller-pixel sensor therefore asks more of the optical system.

However, sampling capability does not guarantee useful image information. If the lens provides very weak MTF near the sensor's highest spatial frequencies, the smallest pixel structure may contain little additional industrial information. The most efficient camera-lens combination is therefore not always the sensor with the smallest possible pixels; it is the combination whose sampling and optical contrast transfer match the actual defect requirement.

Object-Side Resolution Is More Important Than Sensor-Side Numbers for Buyers

The production engineer ultimately needs to know how large one pixel represents on the actual product. If a 1600-pixel horizontal image covers a 320 mm field, the nominal object-side sampling is 320/1600 = 0.20 mm/pixel. A 2 mm defect then spans approximately ten pixels. If the same sensor covers 800 mm, sampling becomes 0.50 mm/pixel, and the same defect spans only four pixels. The camera resolution has not changed, but the amount of spatial information available for that defect has fallen significantly.

This is why the smallest detectable industrial defect cannot be calculated from megapixels without knowing FOV. A buyer should always connect sensor pixels → field of view → millimetres per pixel → pixels across defect → lens contrast at that scale. That chain is far more informative than megapixel count by itself.

Why “Two Pixels Across the Defect” Is Usually a Weak Production Design

Nyquist sampling is sometimes interpreted too literally in machine vision. Two samples may theoretically represent a periodic spatial pattern under ideal conditions, but a production defect is not an ideal sinusoidal target. Its shape may be irregular, its edges may be partially blurred, its SWIR contrast may be weak and its location may fall between sensor pixels. Motion, surface texture and mixed pixels can reduce information further. Designing a system so the minimum defect occupies only two pixels therefore leaves little engineering margin.

A stronger system deliberately places several meaningful pixels across the minimum feature. The exact number depends on the task: presence detection may require fewer than dimensional measurement, while material classification of a small inclusion may require enough interior pixels to avoid host-material mixing. Qualification with representative defects should determine the practical minimum rather than relying on a universal pixel-count rule.

MTF Explains Why Defect Contrast Falls as Features Become Smaller

Imagine a large dark region within brighter material. The transition occupies many pixels, so even moderate optical blur leaves a strong difference. Reduce that region until it becomes comparable with the optical blur diameter, and light from the surrounding material spreads into the defect image while defect signal spreads outward. The dark region becomes lighter, the bright surroundings become slightly darker and the measured contrast decreases.

As the target becomes even smaller, the difference can approach zero. This process is exactly what MTF describes in frequency terms. For industrial inspection, it means the minimum detectable feature is usually reached before the theoretical point at which all optical information disappears because the machine needs enough residual contrast to outperform noise and production variability.

Contrast-to-Noise Ratio Determines Whether Transferred Detail Is Usable

A useful optical defect must survive not only the lens but also sensor noise and normal product variation. If defect pixels have a mean response (μ_D), good-product pixels have mean (μ_G), and their standard deviations are (σ_D) and (σ_G), a simplified contrast-to-noise ratio can be represented as CNR = |μD − μG| / √(σD² + σG²). Two defects with the same nominal intensity difference can therefore have very different detection reliability if one occurs in a highly variable material.

This explains why high MTF is valuable: preserving more optical contrast increases the separation entering the sensor and provides the classifier with greater margin before noise and process variation are considered. For subtle SWIR material inspection, contrast transfer is often more commercially meaningful than visual sharpness.

MTF Should Be Considered at the Wavelength Where the Defect Contrast Exists

A SWIR lens operates over a spectral range rather than one monochromatic wavelength. The inspection system, however, may depend primarily on a smaller spectral region where the target material becomes distinguishable. Optical performance should therefore be considered at the wavelengths relevant to the production decision. A lens that appears extremely sharp under one SWIR band should not automatically be assumed to produce identical contrast transfer throughout the full 900–1700 nm range.

This is particularly important for multi-wavelength systems, material classification and absorption-sensitive inspection. Focus position, diffraction and residual wavelength-dependent aberrations can influence fine-detail contrast. Production qualification should therefore use the illumination bands that the actual machine will use rather than qualifying spatial performance only under an unrelated reference wavelength.

Diffraction Places a Physical Limit on Fine SWIR Detail

As a lens aperture is stopped down, diffraction spreads light from an ideal point into a finite pattern. A commonly used approximation for the Airy-disk diameter is d ≈ 2.44λN, where (λ) is wavelength and (N) is F-number. At 1.5 µm and F4, the approximate Airy diameter is 14.6 µm; at F8, it becomes approximately 29.3 µm. This demonstrates why a very small-pixel sensor can become diffraction-limited when the SWIR lens is stopped down aggressively.

The implication is not that the lens should always remain wide open. Stopping down can improve depth of field and sometimes reduce aberrations, while F1.4 provides stronger light collection and lower diffraction. The correct production aperture is the point where contrast transfer, depth of field, exposure and positional tolerance provide the strongest combined result.

F1.4 Gives the OEM Valuable Optical Throughput

The Kyptec Automation® SWIR portfolio provides an F1.4 maximum aperture across the current family. This can be particularly useful when high-speed inspection requires short exposure or when a narrow SWIR spectral band provides limited illumination energy. More collected light can improve signal-to-noise ratio and help preserve subtle defect contrast.

The engineering decision should still be validated rather than assuming F1.4 is universally optimal. If the inspected product moves significantly in height, a moderately smaller aperture may increase depth of field enough to improve average production MTF because more products remain correctly focused. The best aperture is therefore the one that maximizes usable defect contrast throughout the tolerance range, not necessarily the widest available setting.

Defocus Can Destroy High-Frequency MTF Before an Image Looks Obviously Blurry

One of the most important production effects on fine defect detection is defocus. An operator may consider an image visually acceptable while small-feature contrast has already deteriorated significantly. Large structures remain recognizable because low spatial frequencies survive, but tiny cracks, contaminants or material boundaries can lose contrast much earlier.

This means focus qualification should use the minimum defect rather than only a focus chart or overall image appearance. Move the product through its expected height tolerance, measure defect contrast at each position and identify the point where detection margin begins to collapse. This gives the OEM a practical depth-of-field requirement based on inspection performance rather than subjective sharpness.

Motion Blur Acts Like Another MTF Loss

Movement during exposure reduces contrast along the direction of motion. If conveyor speed is (v) and exposure time is (t), blur distance is approximately b = vt. At 2 m/s and 200 µs, the product moves approximately 0.4 mm during exposure. If the minimum defect is only 0.5 mm wide, this motion can destroy a large fraction of its spatial contrast even when the stationary optical system has excellent MTF.

For this reason, the effective industrial imaging chain should be considered as lens MTF × focus performance × motion transfer × sensor sampling × processing. The lens is fundamental, but final detectability depends on the complete chain. Production validation at maximum line speed is therefore essential when small features are involved.

Wider FOV Reduces Object-Side Spatial Frequency Margin

A broad field is often attractive because one camera can inspect more production area. However, as FOV increases, each physical millimetre receives fewer sensor pixels, moving smaller defects closer to the system's sampling limit. A high-quality lens cannot eliminate this geometric trade-off. If the target occupies only a few pixels, even strong lens-side MTF may not leave enough information for stable classification.

For wide inspection fields, the Kyptec Automation® KL-1408 8.5 MM SWIR Camera Lens offers the broadest focal-length geometry in the current portfolio. The lens can be an effective choice when required defects or material regions remain sufficiently large, but the FOV should be limited by the smallest inspection target rather than the desire to maximize coverage.

Intermediate Focal Length Can Increase Detection Margin Without Excessive Narrowing

When an 8.5 mm field includes unnecessary surroundings, a modest increase in focal length can devote more pixels to the useful product region. The Kyptec Automation® KL-1410 12.5 MM SWIR Camera Lens offers an intermediate option that retains relatively broad coverage while increasing object-side sampling compared with a wider geometry at equivalent working distance.

The important benefit is not simply magnification. Increasing the number of pixels across the target allows the inspection system to operate farther away from the sampling boundary, giving lens MTF, motion and focus losses more margin before the defect becomes unreliable.

A 25 mm Lens Can Be Stronger When the Inspection Region Is Controlled

For individual-product inspection or a limited region of interest, the Kyptec Automation® KL-1412 25 MM SWIR Camera Lens can provide substantially tighter framing than the shorter focal-length options. If the product already occupies a controlled position, using a broad field merely to capture surrounding machine structure wastes sensor sampling that could otherwise be allocated to the defect.

This is particularly relevant when inspection depends on small contamination, narrow cracks, localized coating variation or other fine spatial features. The longer focal length does not change the intrinsic chemistry or spectral contrast of the target; it helps preserve that contrast spatially by providing more samples across the important region.

Longer Focal Length Should Not Be Confused With Higher Lens Resolution

The Kyptec Automation® KL-1414 35 MM SWIR Camera Lens and Kyptec Automation® KL-1416 50 MM SWIR Camera Lens provide progressively narrower fields for appropriate working distances. Their practical advantage for small targets comes from allocating more image pixels to a smaller physical field, not because focal length itself automatically increases optical resolution.

A 50 mm lens used at an unsuitable distance can still create poor inspection geometry, while a shorter focal length can be completely adequate when defects are large. Selection should therefore be driven by required field, minimum feature, available working distance and production margin, not a simplistic assumption that longer means sharper.

Center MTF Is Not Enough for Full-Field Industrial Inspection

Lens performance can vary from the optical axis toward the edge of the image. In addition, illumination falloff, sensor response and focus geometry can make edge positions less favorable than the center. If the production system uses the complete 2/3-inch sensor area, an identical defect should be tested at multiple field positions.

A machine that detects a 1 mm defect reliably in the center but inconsistently near the corner cannot honestly be specified as detecting that defect across the complete field. Qualification should therefore report minimum-defect performance at center, mid-field and required edge locations, with the production claim based on the weakest validated position.

Distortion and MTF Affect Different Aspects of Image Quality

Low distortion and high MTF are both useful, but they solve different problems. Distortion changes where image features appear relative to ideal geometric positions, while MTF determines how well contrast at different spatial frequencies is preserved. A lens can have low distortion yet insufficient high-frequency contrast for a tiny defect, or good fine-detail contrast with geometric distortion that matters for measurement applications.

The Kyptec Automation® SWIR products are positioned for high-contrast, low-distortion industrial imaging across the SWIR range. For buyers, both characteristics should be interpreted according to the application: material classification may tolerate more geometric error than dimensional measurement, whereas microdefect detection may place stronger emphasis on contrast transfer.

Do Not Measure Smallest Detectable Defect From Resolution Charts Alone

Resolution charts are valuable for comparing optical performance, but production targets rarely behave like perfect high-contrast test patterns. A 100 µm dark line on a test chart may be easier to see than a 500 µm moisture anomaly with very low SWIR contrast. The smallest resolvable chart detail should therefore not automatically become the smallest defect claim.

The correct qualification sample is the actual defect—or a representative reference whose contrast, dimensions, material and spectral behaviour closely reproduce it. MTF testing explains the optical system, while real defect testing proves the industrial application.

The Minimum Detectable Defect Should Be Defined Statistically

Instead of reporting that the system “can see 0.5 mm,” an OEM should test multiple examples around the required size and calculate actual detection performance. If 20 defects of approximately 0.5 mm are tested and only 12 are detected reliably, the system has not established dependable 0.5 mm capability. If defects from several lots, field positions and production cycles remain detectable with adequate margin, the claim becomes much stronger.

Minimum defect specification should therefore combine physical size + contrast class + operating conditions + probability of detection. This prevents one unusually easy sample from defining the entire machine specification.

MTF Margin Matters More Than the Absolute Resolution Limit

Operating a production machine at the point where the lens-camera combination barely resolves the minimum target is risky. Small shifts in focus, vibration, illumination or product height can push the system below the detection threshold. A stronger optical architecture places the minimum commercial defect well inside the usable contrast-transfer region.

For example, if a defect theoretically disappears at approximately 5 pixels but becomes stable at 10 pixels, designing the production field around 10–12 pixels offers significantly stronger margin than operating continuously at 5. This engineering reserve is what converts laboratory resolution into robust manufacturing inspection.

Why Kyptec Automation® Is a Strong Platform for Resolution-Critical SWIR Inspection

The Kyptec Automation® SWIR Camera Lens collection provides five focal lengths within the same dedicated 900–1700 nm, 2 MP, 2/3-inch, F1.4 and C-Mount framework. This is valuable for resolution-critical system design because the OEM can adjust object-side sampling through focal length and working-distance choices without leaving the same specialized SWIR optical family. An 8.5 mm configuration can serve broad fields where targets are relatively large, intermediate focal lengths can balance coverage and pixels per defect, and 35 mm or 50 mm options can support tightly framed inspection when the required feature is small and the machine geometry permits additional stand-off.

The practical value of this product range is therefore best understood through system-level resolution. Kyptec Automation® allows buyers to select the field geometry according to the defect requirement while retaining optics designed for the SWIR spectral range where the relevant material contrast is generated.

Frequently Asked Questions About SWIR Camera Lens MTF and Resolution

1. What does MTF tell me that megapixel resolution does not?

Megapixels describe how many samples the camera records, while MTF describes how strongly the lens preserves contrast at different spatial-detail sizes. A sensor may contain enough pixels to represent a defect while the lens transfers too little contrast for reliable detection. For industrial SWIR inspection, both sensor sampling and lens contrast transfer must therefore be considered together.

2. What MTF percentage is required for reliable machine vision?

There is no universal MTF percentage because acceptable contrast depends on the original defect contrast, sensor noise, production variation and classification method. The correct requirement is the MTF level that leaves enough final image contrast for the actual inspection decision. Testing representative defects is therefore more meaningful than applying one universal percentage.

3. Why can a feature occupy several pixels but still disappear?

Pixel coverage alone does not guarantee strong contrast. Optical blur spreads target signal into neighbouring pixels, while surrounding signal spreads into the defect area. Defocus, motion and low material contrast can weaken the difference further. The feature may therefore occupy several pixels geometrically while providing insufficient intensity separation for reliable classification.

4. Is lens MTF more important than camera resolution for small defect detection?

Neither should be considered independently. The camera must sample the feature sufficiently, and the lens must preserve useful contrast at that sampling scale. High MTF with inadequate sensor sampling wastes optical information, while extremely fine sensor sampling with weak optical contrast provides little practical improvement.

5. How does SWIR wavelength affect lens resolution?

Longer wavelengths increase diffraction-related blur for a given F-number, while residual optical performance can also vary across the lens's spectral range. A system using several SWIR bands should therefore verify fine-feature contrast at the actual wavelengths used for inspection rather than assuming resolution is identical throughout 900–1700 nm.

6. Why does stopping down a SWIR lens sometimes reduce fine-detail resolution?

Closing the aperture increases the F-number and therefore increases diffraction blur. Although stopping down may improve depth of field and reduce some aberrations, excessive stopping can reduce high-spatial-frequency contrast. The optimum aperture balances these effects for the actual wavelength and production tolerance.

7. Can F1.4 improve the smallest detectable defect?

It can help by collecting more light and reducing diffraction compared with heavily stopped-down operation, especially when short exposure is required. However, operating at F1.4 can reduce depth of field, so product-height variation may cause defocus. Smallest-defect performance should therefore be qualified across the full mechanical tolerance, not only at perfect focus.

8. Why can the smallest defect be detected in the image center but not near the edge?

Optical contrast, focus and illumination can vary across the field. A fine defect near the edge may therefore have lower contrast-to-noise ratio than the same feature at the center. Full-field validation should place equivalent defects at several positions and define capability from the weakest required location.

9. Is Nyquist frequency the same as the practical resolution of my SWIR system?

No. Nyquist frequency defines a sensor sampling limit based on pixel pitch. Practical resolution is usually lower because lens MTF, diffraction, focus, motion, material contrast and signal-to-noise ratio reduce useful information before the theoretical sampling limit is reached.

10. Can I calculate the minimum defect size using only pixel pitch?

No. Sensor pixel pitch must be converted through magnification or FOV into object-side sampling. You then need to consider how many pixels the target occupies, its SWIR contrast, lens MTF, focus, movement and noise. Pixel pitch alone is a sensor-plane specification, not an object-level detection guarantee.

11. How should I compare two SWIR focal lengths for small-defect inspection?

Calculate the FOV each lens produces at an acceptable working distance and then determine millimetres per pixel and pixels across the minimum defect. Compare whether both options retain enough product-position margin and depth of field. The better lens is the configuration providing sufficient defect sampling without unnecessarily sacrificing required coverage.

12. When is the Kyptec Automation® KL-1408 appropriate despite its wider FOV?

The Kyptec Automation® KL-1408 8.5 MM SWIR Camera Lens is appropriate when broad inspection coverage is required and the smallest relevant features remain large enough in pixels. Wide coverage is valuable for high-throughput material fields, but minimum-defect sampling should always be calculated before final selection.

13. When can the Kyptec Automation® KL-1412 provide a better defect-detection geometry?

The Kyptec Automation® KL-1412 25 MM SWIR Camera Lens can be advantageous when an individual product or smaller region should occupy more of the sensor. The tighter field increases pixels per millimetre compared with a broader configuration at suitable geometry, which can provide more spatial margin for small SWIR features.

14. Does the Kyptec Automation® KL-1416 automatically resolve smaller defects than an 8.5 mm model?

No. The Kyptec Automation® KL-1416 50 MM SWIR Camera Lens provides a narrower FOV for suitable working distances, allowing more sensor pixels to be placed over a smaller region. Actual defect detectability still depends on optical contrast, focus, sensor sampling, wavelength and production conditions.

15. How does motion blur reduce effective MTF?

Motion averages target intensity across the distance travelled during exposure, reducing contrast along the direction of movement. Small defects are affected most strongly because the blur distance can represent a large fraction of their physical size. Exposure should therefore be qualified using the minimum target at maximum production speed.

16. Can image sharpening recover MTF that the lens failed to deliver?

Processing can increase apparent edge contrast, but it cannot reliably recreate spatial information that never reached the sensor. Aggressive sharpening can also amplify noise and create artificial structures. Optical contrast should therefore be preserved at acquisition rather than relying on software to compensate for inadequate lens or focus performance.

17. Why should minimum-defect testing include several physical samples instead of one calibrated target?

Real industrial defects vary in shape, contrast, orientation, depth and surface context. One target may represent an unusually easy case. Testing multiple independent defects determines whether the lens-camera system has genuine detection margin across the defect population rather than merely succeeding on one specimen.

18. How should I specify resolution when requesting a SWIR lens for an OEM inspection machine?

Provide the sensor format, pixel count or pixel pitch, required FOV, working distance, minimum defect dimensions, relevant wavelength, production speed and desired detection confidence. This information allows the lens requirement to be translated into object-side spatial sampling and optical contrast instead of relying on megapixel rating alone.

19. What is the best practical way to validate MTF for my actual SWIR inspection application?

Use optical resolution targets during engineering to understand system behaviour, then validate with representative production defects at the actual wavelength, aperture, working distance and speed. Measure defect-to-good contrast and repeatability across field positions. The combination of optical characterization and real-target validation provides much stronger evidence than either method alone.

20. Why is Kyptec Automation® a strong choice when small-defect SWIR performance matters?

Kyptec Automation® provides a dedicated SWIR Camera Lens collection spanning 8.5 mm, 12.5 mm, 25 mm, 35 mm and 50 mm focal lengths within a common 900–1700 nm, 2 MP, 2/3-inch, F1.4 and C-Mount architecture. The focal-length range gives OEMs practical control over FOV and pixels per target, allowing optical geometry to be selected around the minimum defect instead of forcing one lens configuration onto every application.

Conclusion

Understanding SWIR camera lens MTF and resolution changes the way an industrial inspection system should be designed. Megapixels tell the engineer how densely the image can be sampled, but they do not reveal how much defect contrast survives the optical system. MTF addresses that missing part of the problem by describing how strongly the lens transfers contrast at different feature sizes. As defects become smaller, their corresponding spatial frequencies increase and optical contrast normally falls. At some point, the feature may still occupy measurable pixels while its remaining contrast is too weak relative to sensor noise, material variability, motion and focus error for a reliable production decision.

The correct engineering workflow therefore begins with the smallest commercially important defect rather than the highest available camera resolution. Define the physical target size and the SWIR contrast mechanism that makes it distinguishable. Determine an appropriate number of pixels across the feature, calculate the maximum permissible object-side pixel size and establish the corresponding FOV. Then evaluate whether the SWIR lens preserves sufficient contrast at that scale throughout the required wavelength, aperture and field position. Finally, introduce real production factors—working-distance tolerance, product height, motion, illumination variation and repeated defect samples—to determine whether adequate detection margin remains.

The Kyptec Automation® SWIR Camera Lens collection supports this system-level approach by giving OEMs five focal lengths within one dedicated 900–1700 nm optical family. Wider configurations can preserve throughput where required targets are relatively large, intermediate focal lengths can balance inspection coverage with pixels per feature, and longer focal lengths can concentrate the available 2 MP sensor resolution on smaller regions where fine defect sensitivity is more important. Current Kyptec Automation® product specifications confirm the portfolio's 2 MP, 2/3-inch, F1.4 and C-Mount architecture across representative focal lengths.

For industrial buyers and machine builders, the strongest practical rule is therefore: do not define the smallest detectable defect from megapixels alone—define it from the contrast that remains after the lens, sensor, wavelength, aperture, focus, motion and production variability have all acted on the target. When optical contrast transfer and object-side sampling are engineered together, Kyptec Automation® SWIR Camera Lenses provide a strong foundation for building 900–1700 nm inspection systems in which resolution is measured by reliable defect detection rather than by specification-sheet numbers alone.