Fluorescence inspection lighting: excitation and filters

A transparent adhesive bead, clear coating or faint residue can disappear under ordinary reflected light. Fluorescence can create useful contrast, but only when the target emits differently from the surrounding material. Reflected excitation, background fluorescence and ambient light can otherwise overwhelm the feature.

This guide helps you evaluate samples before selecting hardware. Start with controlled excitation and emission-filter trials, then calculate sampling and exposure requirements. The goal is a repeatable inspection image—not merely a bright glow during a stationary demonstration.

Decide whether fluorescence answers the inspection question

Fluorescence occurs when a material absorbs excitation light and emits light, usually at longer wavelengths. The camera should receive the useful emission while rejecting reflected excitation. That separation is the main difference from ordinary color or reflectance inspection.

A fluorescent image might reveal adhesive presence, coating coverage, a printed marker or contamination. It does not automatically measure thickness, bond strength or chemical identity. Those conclusions require separate validation against reference measurements.

Choose the method according to the physical difference you need to detect:

For reflectance-based alternatives, review how LED color changes inspection contrast. If the task requires material identification rather than presence detection, use the distinctions in spectral fingerprint identification.

Prove the effect on representative samples

Begin with known acceptable and unacceptable parts. Include missing material, minimum coverage, excess material and likely contamination where those conditions matter. A perfectly coated sample alone cannot establish an inspection limit.

For an initial screening trial, a practical starting set is 10 acceptable and 10 unacceptable samples from several production conditions. This is a feasibility screen, not a statistical acceptance plan. Expand the sample set before approving production performance.

Use the following sequence:

1. Fix the sample, camera, lens, aperture and working distance.
2. Disable automatic exposure, gain, white balance and tonal adjustments where available.
3. Record an image with the excitation off to measure ambient contribution.
4. Illuminate the sample with a candidate excitation band using appropriate optical shielding.
5. Install a candidate emission filter and repeat without changing the sample position.
6. Capture at least 3 repeats per condition and save the acquisition settings.

Record the bare substrate, target material and a known non-target residue separately. Otherwise, an image of a bright coated area might actually be showing substrate fluorescence or reflected light.

Select excitation by contrast, not by the UV label

UV is one possible excitation region, not a universal requirement. Some fluorescent materials respond to violet or blue excitation. The useful band depends on the material's absorption and the contrast against its background.

Test candidate bands using the same imaging geometry. Record source distance, exposure, aperture and filter arrangement. When comparing excitation efficiency, account for differences in incident irradiance; equal controller settings do not imply equal optical output.

Ask for the source spectrum, not only its nominal wavelength. Spectral tails can overlap the emission passband and appear as reflected leakage. Also check illumination uniformity across the inspection area, particularly where coating coverage is judged by brightness.

The special lights category and lighting finder are starting points for sourcing. A category label alone does not establish a suitable excitation spectrum, output level or compatible emission filter.

If several materials need different excitation bands, capture and evaluate them separately. The guide to sequencing multiple illumination colors explains the acquisition approach, but each fluorescent material still needs its own contrast trial.

Choose filters to reject excitation and pass emission

The emission filter sits in the camera's optical path. Its job is to transmit useful fluorescence while suppressing the excitation reaching the sensor. An excitation filter may also be needed at the source to remove unwanted wavelengths.

Two common emission-filter approaches have different trade-offs:

• Long-pass: transmits wavelengths above an edge. This can collect a broad emission band, but also admits more ambient light and unrelated fluorescence.
• Band-pass: transmits a bounded wavelength range. This can isolate a useful emission region, but usually collects less total light.

Do not select either type from its name alone. Request transmission curves across the excitation and emission regions, the specified blocking range, and the conditions under which those curves apply.

Interference-filter passbands can shift with incidence angle. A filter that works in a narrow, near-normal beam may behave differently across a wide-angle lens. Check the intended mounting location, clear aperture and angular limits.

Use lens accessories and the accessories finder to identify mechanical options after defining the optical requirement. Thread compatibility does not establish spectral compatibility.

The lens normally needs to transmit the emitted wavelengths, not the excitation wavelengths that you intend to block. A visible-emission application therefore does not automatically require a UV-transmitting lens. The UV and SWIR lens basics guide explains when imaging outside the visible range changes lens selection.

Estimate blocking and camera sampling

Excitation leakage can dominate a weak fluorescent signal even when the reflected light looks dim to the eye. Optical density expresses filter attenuation at a specified wavelength:

For a simplified estimate, suppose the unblocked excitation contribution at the detector would be 10,000 times the useful fluorescence signal after emission filtering. You want leakage below 10% of that useful signal:

This calculation gives a blocking target at the excitation wavelengths, not a complete filter specification. Source bandwidth, angular effects, stray light and fluorescence transmission still matter. Do not infer blocking performance from a filter's center wavelength.

Next, size the image around the smallest feature. Suppose the field of view is 100 mm wide and the smallest missing-coating gap is 0.20 mm. Using 4 pixels across that gap as an initial sampling target:

Four pixels across a feature is a starting assumption, not a guaranteed detection limit. Optical blur, emission contrast, processing and acceptable error rates determine the final requirement.

Validate the smallest defect near image corners as well as the center. Illumination falloff or filter angle effects may make an edge location the limiting case.

Choose the camera after measuring the emission

A monochrome camera is often the first candidate when the decision depends on brightness within one emission band. A color camera becomes useful when the emission color itself provides information that survives filtering and production variation.

Neither choice establishes sensitivity at the required wavelengths. Obtain spectral-response information and check the complete optical chain before ordering. Interface type and pixel count cannot replace that check.

The MV-CS200-10UM is a monochrome USB3 camera with a rolling shutter. It is a candidate for evaluating monochrome acquisition, subject to spectral-response and timing checks.

The MV-CU020-90GC is a color GigE camera with a global shutter. Consider this acquisition approach when motion handling or emission-color information matters, then verify sensitivity and image quality on your samples.

These are camera candidates, not a validated excitation-and-filter package. Compare the area scan cameras through the camera finder after defining the needed sampling, exposure, interface and shutter behavior.

For a rolling-shutter camera, verify row timing and illumination compatibility explicitly. A short light pulse does not necessarily expose every row equally. A stationary demonstration cannot establish suitability for a moving line.

Separate exposure limits from ambient-light problems

Start exposure trials on a stationary sample. For example, capture at 0.5 ms, 1 ms, 2 ms and 4 ms while holding gain and illumination constant. Check the target and background distributions, not just the displayed image.

Keep both target and background below saturation. If increasing exposure makes both equally brighter without improving their separation, more exposure is not solving the underlying contrast problem.

Then apply the motion constraint. Using the earlier object sampling, a line speed of 1,000 mm/s and an allowable blur of 2 pixels:

A stationary result obtained at 2 ms therefore needs a much shorter effective exposure on this moving line. Under a linear-response assumption, collecting the same signal in 100 microseconds requires about 20 times the detected photon rate. That is a feasibility warning, not permission to exceed a light's ratings.

Review exposure and shutter timing for moving parts before choosing the acquisition method. If pulsed illumination is needed, follow documented electrical and thermal limits as described in strobe duty cycle and timing.

Measure ambient sensitivity separately. Capture excitation-off images under the brightest expected room-light condition, then repeat with optical screening. An emission filter cannot reject ambient light that lies inside its own passband.

Verify separation before setting a threshold

Use raw or consistently processed image values. Automatic display scaling can make two captures look equally clear despite very different signal levels.

Measure a target region and a background region with comparable geometry. One useful screening metric is:

For example, target and background means of 120 and 70 counts, each with a standard deviation of 5 counts, give:

This metric helps compare trial settings. It does not prove a production error rate, especially when images contain gradients, correlated pixels or mixed defect types. Evaluate part-level measurements across the full sample set before setting acceptance limits.

Repeat the measurement after realistic position changes, material-lot changes and permitted ambient variation. Check whether prolonged illumination changes the response through photobleaching or material heating. Do not assume a stable stationary sample represents repeated production exposure.

Recognize when fluorescence is the wrong method

Fluorescence is a poor choice when the target and substrate emit similarly, or when the process does not control the fluorescent ingredient. A bright adhesive may still be incorrectly cured or poorly bonded.

Brightness is also not automatically proportional to coating thickness. Excitation absorption, emission reabsorption, concentration, viewing geometry and saturation can make the relationship nonlinear. Establish any thickness calibration against independent measurements over the actual process range.

Reject the approach if the required motion-limited signal cannot be achieved within safe, documented operating limits. Consider reflected-light inspection, a different tracer approved for the process, or another measurement method instead.

Do not keep adding gain to compensate for leakage or background fluorescence. Gain changes recorded levels; it does not create the missing optical separation.

Check the specification before ordering

Use this checklist to turn a successful trial into a purchasing specification:

• Inspection claim: define presence, coverage, gap size or another observable result. Avoid substituting fluorescence brightness for an unverified material property.
• Sample range: include acceptable variation, known defects and likely background contaminants. Save representative images and measurement distributions.
• Excitation: record wavelength band, source spectrum, geometry and required irradiance. Confirm the operating limits used during the trial.
• Emission filtering: specify passband, blocking range, required attenuation, clear aperture and mounting-angle limits.
• Camera and lens: confirm spectral response, field of view, sampling, focus and shutter behavior. Recheck image corners with the final filter fitted.
• Timing: document line speed, blur allowance, exposure and trigger sequence. Verify the complete sequence at production speed.
• Ambient and safety: define optical screening and safe access arrangements. Recheck contrast under the brightest permitted ambient condition.
• Acceptance test: agree on reference samples and pass/fail measurements before committing to the installation.

Start the lighting configuration with the measured excitation and emission requirements. For help selecting a matched optical chain, contact applications support with sample images, filter curves, field of view, smallest defect and motion-limited exposure.