SWIR Camera Lens for Liquid and Chemical Concentration Inspection: Detecting Composition, Dilution and Process Variation

Industrial liquid inspection becomes difficult when the quality difference that matters is chemical rather than visual. Two solutions may appear identical in colour and transparency while containing different concentrations of water, solvent, dissolved solids, oils, active ingredients or other constituents. A mixture can also drift gradually during production without developing any visible defect that a conventional imaging system can reliably detect. This is where 900–1700 nm SWIR imaging for liquid and chemical concentration inspection becomes valuable. Short-wave infrared imaging can exploit wavelength-dependent absorption and transmission differences within liquids and chemical mixtures, allowing an industrial vision system to detect changes that are associated with composition, dilution, mixing consistency, concentration variation or process drift. The important engineering principle is that a SWIR camera does not directly “read concentration” simply because it operates in the infrared; useful concentration measurement depends on selecting spectral regions where the composition changes the optical response, then controlling optical path length, illumination, lens geometry, sensor exposure, reference conditions and calibration closely enough that composition remains the dominant source of measured variation.

The dedicated Kyptec Automation® SWIR Camera Lens collection provides 8.5 mm, 12.5 mm, 25 mm, 35 mm and 50 mm focal lengths for 900–1700 nm imaging, with the current portfolio specified around 2 MP resolution, 2/3-inch sensor format, F1.4 aperture and C-Mount. This gives OEMs and machine builders several practical optical geometries for inspecting liquid streams, transparent vessels, process cells, filling lines, chemical containers and controlled regions of interest. The lens does not create chemical selectivity by itself, but it plays a fundamental role in ensuring that the material region is represented with enough spatial purity, optical throughput and repeatable geometry for spectral differences to become reliable production information.

Why Liquid Composition Can Produce Useful SWIR Contrast

Liquids interact with short-wave infrared radiation according to their molecular composition, concentration, optical path length and physical condition. At some wavelengths, a liquid may transmit a relatively large proportion of the incident energy; at others, molecular absorption can reduce the detected signal substantially. If a dissolved component or mixture ratio changes the absorption behaviour within the selected spectral region, the camera can record a corresponding intensity change. This creates the basis for SWIR liquid concentration measurement, chemical composition inspection, dilution detection and inline process monitoring.

The important distinction is between detecting a spectral response and calculating an absolute concentration. A machine may easily distinguish an acceptable formulation from a severely diluted formulation while still lacking the calibration accuracy required to report an exact chemical percentage. Industrial system design should therefore begin by deciding whether the requirement is binary verification, grade classification, process-trend monitoring or quantitative concentration estimation. Each objective requires a different level of calibration, spectral information and reference control.

Concentration Measurement Starts With the Optical Path Through the Liquid

The amount of SWIR energy absorbed by a liquid depends not only on composition but also on how far the radiation travels through the sample. A 2 mm liquid layer and a 20 mm liquid layer of exactly the same formulation can produce very different measured intensities. For many absorbing systems, the general relationship resembles Beer-Lambert behaviour, where transmitted intensity falls as concentration and optical path length increase. A simplified conceptual form is:

A = ε × c × L

where A represents absorbance, ε represents an absorption coefficient associated with the wavelength and material, c is concentration and L is optical path length.

Industrial liquids are often more complicated than this simplified model because scattering, multiple constituents, container surfaces, temperature and detector response can also influence the measurement. The formula nevertheless illustrates why a concentration inspection station should control the liquid thickness or measurement path. If the path length changes substantially while concentration remains constant, the system can incorrectly interpret geometry as chemistry.

Transmission Imaging Can Be Powerful When the Liquid and Container Allow It

In a transmission arrangement, SWIR illumination passes through the liquid before reaching the camera. This can provide strong sensitivity to absorption because the detected signal has interacted with a known volume of material. For concentration or dilution inspection, transmission can be especially useful where the container, process window and liquid remain sufficiently transmissive at the selected wavelengths.

The design should define a repeatable path length and avoid optical saturation at both extremes. If absorption is too weak, concentrated and diluted samples can produce nearly identical signals. If absorption is too strong, almost no signal reaches the sensor and different concentrations collapse toward the same dark value. The optimum wavelength and path length therefore create a measurable response curve between these extremes rather than maximizing absorption blindly.

Reflectance Imaging Can Be Better When Transmission Is Not Practical

Some industrial liquids are inspected in open vessels, opaque containers, coatings, flowing films or process surfaces where backlighting is impossible. In these cases, reflected SWIR energy may carry useful information about composition. The illumination and camera observe radiation returned from the liquid surface and near-surface volume rather than energy transmitted through the entire sample.

Reflectance systems require careful geometry because surface reflections can become large compared with the composition-sensitive signal. Glossy liquids can generate bright specular highlights that change dramatically with viewing angle. A robust machine should therefore stabilize illumination angle, camera angle, liquid level and surface motion so that the material response does not become confused with changing reflection geometry.

Dilution Detection Is Often Easier Than Exact Concentration Measurement

Many production lines do not need laboratory-grade concentration values. They need to know whether a formulation has been diluted beyond an acceptable limit, whether the wrong mixture entered a process, or whether a batch is moving away from its approved recipe. These are often easier SWIR problems because the system can be trained or calibrated against acceptable and unacceptable populations rather than estimating an exact percentage.

For example, if an approved solution produces a normalized SWIR response between 0.58 and 0.64 while clearly diluted material produces 0.39–0.45 under controlled conditions, the machine can establish a decision region between the populations. The actual numbers depend entirely on the material, wavelength and optical geometry, but the principle is broadly applicable: a strong industrial measurement needs separation between populations, not merely a visible intensity difference between two convenient samples.

Multi-Wavelength Ratios Can Separate Concentration Change From Illumination Drift

Absolute intensity can change when illumination output, container position, surface condition or camera exposure varies. A multi-wavelength architecture can sometimes improve concentration robustness by comparing a composition-sensitive band against a less-sensitive reference band. If (I_s) is the sensitive wavelength response and (I_r) is a reference wavelength, a simple ratio can be calculated as:

R = Iₛ / Iᵣ

A normalized difference can also be used:

ND = (Iₛ − Iᵣ) / (Iₛ + Iᵣ)

The purpose is to create a feature that responds more strongly to composition than to common brightness changes. Ratios do not automatically solve every stability problem; they still require adequate signal, non-saturated images and repeatable spatial registration. They can, however, make liquid and chemical inspection substantially more robust when a single raw intensity value is influenced by nuisance variables.

Calibration Should Use the Full Expected Concentration Range

A concentration model developed from only a “good” sample and a “bad” sample may behave unpredictably between those points. If the production requirement involves quantitative or semi-quantitative concentration, calibration samples should cover the complete operating range with enough intermediate levels to reveal whether the optical response is linear, nonlinear or saturated.

Suppose reference solutions are prepared at 0%, 5%, 10%, 15%, 20% and 25% concentration. Their SWIR response may decrease nearly linearly over part of the range and then flatten because absorption becomes too strong. If the production limit lies inside the flattened region, a different wavelength or shorter optical path may provide greater sensitivity. Calibration therefore informs optical design rather than merely producing a software equation after the hardware is fixed.

The Lens Must Give the Liquid Region Enough Pixels for Stable Measurement

When a chemical measurement is made from a region of interest, the liquid should occupy enough pixels that the calculated intensity is not dominated by container edges, bubbles, labels, meniscus curvature or background. A region containing only a few pixels can change substantially when the container moves by a small amount. A larger, clean interior measurement region allows spatial averaging and produces more stable concentration features.

For broad process or filling-line coverage, the Kyptec Automation® KL-1408 8.5 MM SWIR Camera Lens can be evaluated where multiple containers or a wider liquid area must be observed within limited working distance. The current model is specified for 900–1700 nm, 2 MP, 2/3-inch format, F1.4 and C-Mount. The wide field is useful only when each required liquid region still receives enough sensor pixels to support stable spectral averaging.

Container Material Must Be Treated as Part of the Optical Measurement

When a liquid is inspected through a bottle, tube, sight glass or process window, the container material contributes its own wavelength-dependent transmission and reflection. A container that appears perfectly transparent to the human eye may not behave uniformly throughout 900–1700 nm. Its wall thickness may also vary across the field, and curved surfaces can alter optical path length.

A strong calibration should therefore use the same production container or window used in the machine. If several approved container suppliers are used, each should be included in validation. The system should determine whether the container response is stable enough to normalize or whether it introduces unacceptable variation near the chemical decision boundary.

Curved Bottles Can Create False Concentration Gradients

A cylindrical container makes the optical path through the liquid longer near some viewing positions and shorter near others. It can also refract light so that different pixels observe different effective path lengths. As a result, the image may show a brightness gradient even though the concentration is perfectly uniform.

Instead of averaging the entire bottle indiscriminately, the inspection can use a carefully selected central region where wall geometry and path length are more predictable. Mechanical guides can also stabilize lateral bottle position. For high-accuracy applications, an engineered measurement cell with controlled path length may be preferable to inspecting through arbitrary consumer packaging.

Liquid Level Variation Must Not Be Confused With Concentration Variation

If the measurement region moves close to the liquid-air boundary, the signal can change because of the meniscus, reflections, changing optical path or partial pixels containing both liquid and air. This can look like a concentration change to an algorithm. The solution is to separate fill-level and concentration measurements geometrically.

The concentration region of interest should remain fully inside the liquid across the permitted fill tolerance. If the liquid level itself is highly variable, the software should first locate the stable liquid region and then calculate spectral features only from valid pixels.

Bubbles and Foam Can Produce Large SWIR Measurement Errors

Bubbles replace liquid with gas and introduce additional reflective interfaces, while foam creates a highly scattering surface structure. Both can substantially change measured intensity without any real change in chemical concentration. A production system should determine whether bubbles are an abnormal condition that should itself trigger rejection or a normal process state that must be filtered from the concentration measurement.

Spatial statistics can help. If most of a measurement region is uniform but a small number of pixels show extreme values associated with bubbles, robust median or percentile-based features may outperform a simple arithmetic mean. If bubbles are widespread, the machine may need to delay inspection until the fluid settles or move the measurement location to a more stable section of the process.

Mixing Uniformity Can Be Inspected Spatially Rather Than From One Average Value

A liquid batch can have the correct average composition while still being incompletely mixed. If one side of a process cell contains more concentrated material and another side contains more diluted material, averaging the whole image may hide the problem. One advantage of imaging over a single-point detector is the ability to examine spatial concentration uniformity.

The field can be divided into multiple regions, each producing its own normalized spectral feature. The system can then evaluate both mean composition and variation across the field. A large spatial standard deviation can indicate incomplete mixing even when the overall average appears correct.

The Kyptec Automation® KL-1410 Can Support Mid-Field Liquid Inspection

The Kyptec Automation® KL-1410 12.5 MM SWIR Camera Lens provides a useful intermediate focal length for systems that need broader coverage than a product-focused inspection but do not need the maximum field of the 8.5 mm option. Its current specification includes 12.5 mm focal length, 900–1700 nm operation, 2 MP resolution, F1.4 aperture, 2/3-inch sensor format and C-Mount. In liquid inspection, this geometry can be valuable when a machine must cover several process regions while still preserving enough pixels within each measurement zone for stable concentration analysis.

Temperature Should Be Included in Chemical Calibration When It Affects the Process

Temperature can change liquid density, optical properties and physical behaviour, while temperature variation may also correlate with process stage. If calibration samples are measured at one temperature but production material spans a much wider range, the concentration model can drift even if composition is unchanged.

The practical solution is to test the expected temperature envelope during qualification. If temperature has a measurable effect, the machine can either constrain the measurement to a stable thermal condition, build temperature compensation into the model or establish separate calibration curves. Ignoring a known process variable is usually weaker than explicitly incorporating it into the measurement architecture.

Flowing Liquids Require Exposure and Motion to Be Considered Differently

For a perfectly homogeneous liquid, motion during exposure may have little effect because neighbouring regions contain essentially the same composition. For mixed streams, suspended material, concentration fronts or incomplete mixing, however, motion blur can average physically different regions together. The exposure time should therefore reflect the smallest spatial concentration variation the system needs to preserve.

The F1.4 maximum aperture available across the current Kyptec Automation® SWIR range provides useful light-gathering capability for short-exposure applications. The final aperture should still balance signal, focus tolerance and optical performance according to the actual process geometry rather than remaining permanently wide open by default.

A 25 mm SWIR Lens Can Improve Measurement of Controlled Liquid Cells

Where concentration is measured through a small, controlled process cell rather than a broad production line, tighter framing can make better use of the available sensor area. The Kyptec Automation® KL-1412 25 MM SWIR Camera Lens can be evaluated when the liquid measurement zone should occupy a larger proportion of the image. By reducing irrelevant background, a narrower FOV can support larger interior regions of interest and reduce sensitivity to adjacent structures.

This is especially useful when concentration differences are subtle and the processing method benefits from averaging many clean pixels from one controlled liquid path.

Concentration Thresholds Should Include an Uncertain Zone

If acceptable and unacceptable concentration distributions approach each other near a specification limit, forcing every measurement into pass or fail can make the process unstable. An uncertainty zone can be used around the decision boundary. Products inside this region may be remeasured, routed for secondary testing or handled according to the application's quality-control strategy.

For example, an optical feature below 0.42 might be accepted, values above 0.48 rejected and values from 0.42 to 0.48 marked uncertain. The actual values must come from application-specific calibration, but the principle helps prevent borderline optical noise from generating unjustified confidence.

Process Drift Is Often More Valuable to Detect Than a Single Out-of-Spec Sample

One of the strongest uses of SWIR concentration imaging is continuous process monitoring. A production line may remain technically within tolerance while its spectral response gradually trends toward the reject boundary. Detecting that direction early can allow an operator or control system to correct the process before significant scrap is produced.

A control chart can track normalized SWIR response over time. If the acceptable historical mean is 0.57 and consecutive batches move through 0.56, 0.54, 0.52 and 0.50, the trend itself may be important even before the formal rejection threshold is crossed. SWIR imaging can therefore become part of process variation detection, not merely end-of-line sorting.

Longer Focal Lengths Can Support Remote or Localized Chemical Inspection

Some chemical-process environments require the optical hardware to remain farther from the liquid because of mechanical equipment, enclosure design or process access. The Kyptec Automation® KL-1414 35 MM SWIR Camera Lens and Kyptec Automation® KL-1416 50 MM SWIR Camera Lens provide narrower focal-length options where a small liquid region must be framed from greater stand-off. The advantage comes from geometry and sensor utilization, not from any assumption that longer focal lengths are inherently more chemically sensitive.

Validation Must Use Independent Production Samples

The same samples used to create a concentration model should not be the only samples used to prove it. Independent lots should be tested after the calibration equation or classification threshold is fixed. This is particularly important for chemicals whose impurities, supplier variation or process history may change optical response slightly.

A strong validation set includes multiple concentrations around the critical decision boundary, several production lots, expected temperature variation, container tolerances, different positions in the field and repeated measurements. The final performance claim should be based on this independent population rather than calibration fit alone.

Calibration Accuracy and Classification Accuracy Are Different

A regression model may estimate concentration with an average error of ±1%, yet still perform poorly at a critical pass/fail threshold if errors are concentrated near that boundary. Conversely, a system may classify “correct formulation versus diluted formulation” almost perfectly while providing poor absolute concentration values.

The buyer should therefore define the metric according to the production requirement. If the machine only needs to prevent wrong formulation from entering production, classification accuracy may matter more than precise concentration measurement. If the machine controls chemical dosing continuously, quantitative error and repeatability become more important.

Why Kyptec Automation® Is a Strong Optical Platform for Liquid and Chemical SWIR Inspection

The Kyptec Automation® SWIR Camera Lens collection provides five focal lengths spanning 8.5 mm, 12.5 mm, 25 mm, 35 mm and 50 mm within a dedicated 900–1700 nm lens family. Current product pages verify key specifications including 2 MP resolution, 2/3-inch sensor format, F1.4 aperture and C-Mount. This range is valuable for liquid inspection because process geometry can vary dramatically: one machine may need a broad field covering several containers, another may examine a narrow flow cell, and another may require significant stand-off from a chemical process. Kyptec Automation® gives OEMs the flexibility to select focal length according to the optical measurement zone while remaining within the same SWIR-focused product category.

Frequently Asked Questions About SWIR Liquid and Chemical Concentration Inspection

1. Can SWIR imaging measure liquid concentration without touching the liquid?

Potentially, yes. If the liquid produces a concentration-dependent optical response within the chosen SWIR wavelengths and the light can interact with the liquid through a stable optical path, non-contact imaging can estimate or classify concentration. The system still requires calibration against known reference samples and must control path length, illumination, container effects and temperature sufficiently for the required accuracy.

2. Can SWIR distinguish a correct chemical mixture from an over-diluted mixture?

Yes when dilution changes the spectral response enough to create repeatable separation. The strongest design tests several dilution levels around the actual specification limit rather than comparing only extremely different samples. A threshold or multiband feature can then be created from the production population.

3. Does a darker SWIR image always mean a higher chemical concentration?

No. Image darkness can be affected by wavelength, path length, illumination, container transmission, material absorption and camera exposure. Higher concentration may increase absorption at a useful wavelength, but the relationship should be established experimentally rather than inferred from brightness alone.

4. Why does the thickness of a liquid sample affect concentration measurement?

A longer optical path gives SWIR radiation more material through which to travel, which can increase absorption even when concentration remains unchanged. If path length varies, the system can confuse geometry with chemistry. Controlled measurement cells or carefully selected bottle regions can reduce this problem.

5. Can SWIR inspect concentration through a transparent bottle?

It can when the bottle material transmits enough energy at the selected wavelengths and its wall properties remain sufficiently consistent. The production container must be included in calibration because visible transparency does not guarantee uniform transmission between 900 and 1700 nm.

6. How do bubbles affect a SWIR concentration measurement?

Bubbles replace part of the liquid path with gas and introduce reflective interfaces, so they can change intensity substantially. The system may need to identify bubbles and exclude them from the measurement region, delay inspection until the fluid settles, or reject excessive bubbling as a separate process condition.

7. Can SWIR detect whether two clear liquids have different compositions?

Potentially, even when they appear identical in visible light, because visually transparent liquids can have different absorption behaviour in the SWIR spectrum. Actual separability depends on the chemistry, wavelength range, path length and required accuracy, so representative testing is essential.

8. Is one SWIR wavelength enough for concentration inspection?

Sometimes. If one wavelength produces stable separation across the complete concentration and process range, a single-band system can be simpler and faster. If illumination, container variation or overlapping responses reduce reliability, a second reference wavelength can improve normalization or class separation.

9. Can SWIR imaging detect incomplete mixing inside a liquid batch?

Yes when different mixture concentrations produce measurable spectral contrast and the camera has a spatial view of the liquid. Instead of calculating one global average, the image can be divided into multiple regions so the system detects concentration gradients or unusually high spatial variation associated with incomplete mixing.

10. How should I calibrate a SWIR system for liquid concentration?

Prepare traceable reference samples covering the full expected operating range, including several points around the critical specification limits. Capture them under the final illumination, path length, temperature, container and lens geometry. Build the calibration from those samples and validate it using independent material not included in model fitting.

11. Can SWIR concentration inspection replace laboratory chemical analysis completely?

Not automatically. SWIR can be highly valuable for rapid inline screening, classification and process monitoring, but whether it can replace a laboratory method depends on the chemical system, required uncertainty, regulatory requirements and validation results. Many applications use SWIR to inspect every product or continuously monitor a process while laboratory methods remain the reference for calibration.

12. When is the Kyptec Automation® KL-1408 useful for liquid inspection?

The Kyptec Automation® KL-1408 8.5 MM SWIR Camera Lens can be considered when a wide process area, several containers or a broad liquid surface must fit into the inspection field. The OEM should still ensure that every concentration measurement region contains enough clean pixels and is not dominated by container edges or background.

13. When can the Kyptec Automation® KL-1412 be useful for chemical concentration inspection?

The Kyptec Automation® KL-1412 25 MM SWIR Camera Lens can be useful when the measurement is concentrated on a controlled flow cell, bottle region or process window. Tighter framing can devote more of the 2 MP sensor to the useful liquid area and improve the stability of region-based spectral measurements.

14. Why does my concentration result change when the bottle moves sideways?

Lateral movement can change the effective wall thickness, curvature, optical path and measurement region. It may also move the region into a different illumination level. Mechanical guides, image-based region tracking and reference normalization can reduce these position-related errors.

15. Can a SWIR system detect very small concentration changes?

Potentially, but the achievable sensitivity depends on the strength of the spectral change relative to noise and process variability. Small concentration differences require stable illumination, controlled optical path, adequate sensor signal, carefully chosen wavelengths and enough calibration samples near the required measurement range.

16. How do I know whether concentration should be measured using transmission or reflection?

Transmission is attractive when light can pass through a controlled thickness of liquid and provides measurable absorption without becoming too weak. Reflection is useful where backlighting is physically impossible or the liquid is inspected from one side. The choice should be based on the material, container, process access and spectral contrast available in each geometry.

17. Why should concentration calibration include temperature?

Temperature can alter physical and optical properties of the liquid and may change density, viscosity or spectral response. If production temperature varies significantly, calibrating at only one temperature can introduce systematic measurement error. Testing the expected temperature range reveals whether compensation or tighter process control is required.

18. Can a SWIR camera detect process drift before the liquid becomes out of specification?

Yes. Continuous normalized SWIR measurements can reveal gradual movement in material response before the formal reject threshold is crossed. Trending these measurements over time can support early process intervention and reduce the risk of producing a large quantity of off-spec material.

19. What information should I provide before selecting a SWIR lens for liquid or chemical inspection?

Provide the liquid or formulation being inspected, target concentration range, required pass/fail limit or quantitative accuracy, selected wavelengths if known, container or process-window material, optical path length, active sensor size, FOV, working distance, process speed and available mounting space. These inputs allow the SWIR lens geometry to be matched to the actual measurement task rather than selecting focal length independently.

20. Why is Kyptec Automation® a strong choice for SWIR liquid and chemical inspection systems?

Kyptec Automation® provides a dedicated SWIR Camera Lens collection covering 8.5 mm, 12.5 mm, 25 mm, 35 mm and 50 mm focal lengths within a 900–1700 nm, 2 MP, 2/3-inch, F1.4 and C-Mount optical family. That focal-length breadth allows OEMs to design wide filling-line inspection, medium-field process monitoring, controlled flow-cell analysis or narrow stand-off measurement while staying within a consistent SWIR-focused lens portfolio. The result is a practical optical platform for turning composition-sensitive SWIR contrast into repeatable industrial measurements when the material and process have been properly validated.

Conclusion

Using a SWIR camera lens for liquid and chemical concentration inspection requires the imaging system to distinguish real composition changes from every other factor that can change detected intensity. Concentration, dilution and mixture ratio can influence absorption within the 900–1700 nm spectrum, but so can optical path length, container thickness, surface reflections, bubbles, temperature, illumination variation and product position. A reliable machine therefore cannot be designed around the assumption that darker automatically means more concentrated or that one attractive SWIR image proves quantitative capability. The optical measurement must be engineered so that composition remains the strongest and most repeatable source of variation.

The strongest workflow begins with representative materials at known concentrations. Candidate SWIR wavelengths should be screened across the complete process range, including intermediate and boundary conditions. The engineer should then decide whether the production task requires exact concentration estimation, pass/fail formulation verification, dilution detection, mixture uniformity inspection or process-trend monitoring. Optical path length should be stabilized, container effects characterized, exposure optimized and the liquid region assigned enough sensor pixels to support robust averaging. Where absolute intensity remains vulnerable to common brightness changes, multi-wavelength ratios or normalized measurements can provide stronger material-specific information.

The Kyptec Automation® SWIR Camera Lens collection provides a strong foundation for building these systems because its five focal lengths allow the physical inspection geometry to be optimized independently of the concentration model. The Kyptec Automation® KL-1408 can support broad fields when several vessels or a wide process region must be monitored, intermediate focal lengths can balance coverage with region-of-interest size, and longer options can dedicate more of the sensor to a controlled liquid cell or allow useful stand-off from process equipment. Across the current portfolio, Kyptec Automation® maintains a focused 900–1700 nm, 2 MP, 2/3-inch, F1.4 and C-Mount architecture for industrial SWIR integration.

For OEMs and industrial buyers, the most important principle is to treat SWIR concentration imaging as a controlled optical measurement rather than simple infrared photography. The selected wavelength must respond meaningfully to composition, the optical path must remain known, the lens must isolate a clean and sufficiently sampled material region, and calibration must survive real production variables rather than only laboratory conditions. When these requirements are engineered together, Kyptec Automation® SWIR Camera Lenses provide a flexible and technically appropriate optical platform for non-contact liquid composition verification, dilution detection, mixture monitoring and process-variation inspection across demanding 900–1700 nm industrial environments.