Changing Machine Vision Lighting: Geometry, Wavelength and Uniformity to Verify

Illumination is the first optical element

Lens changes get the attention; lighting changes cause more inspection failures. The illumination determines which features produce contrast at all. Change it and a defect mode that was clearly visible can disappear entirely, while the lens and camera are untouched.

Across 2,602 machine vision companies in the POMEAS global intelligence database, Moritex appears in 5 company profiles as a named lighting supplier. The count is low compared with camera and lens brands, which reflects how lighting is specified — usually as part of a station rather than as a named component — not how much it matters.

This article lists what to verify before substituting illumination. It is a verification checklist, not a cross-reference. Nothing here states or implies that a POMEAS light replaces any third-party model.

1. Geometry: which geometry reveals your feature

Geometry is the first decision because it determines everything else.

POMEAS publishes these as distinct series: GH (LED ring, multiple sizes, red/white/blue), GHN (diffuse shadowless ring), GL (LED bar, large-field uniform), GBT (parallel surface), VP-GLTD (telecentric parallel backlight, 30–265 mm) and VP-GLTDL (right-angle telecentric backlight, 35–265 mm).

What to verify: which geometry currently makes your defect visible. Replicating the geometry matters more than matching brightness.

2. Wavelength: colour is a contrast tool, not a preference

Wavelength selection changes contrast between materials:

- Red — longest wavelength, diffracts least,good for general use and for penetrating some substrates - Blue — shorter wavelength, better scattering detail on metal surfaces, higher sensitivity to fine texture - White — broad spectrum, use when colour is part of the inspection - UV (365 / 395 nm) — excites fluorescence; used for UV adhesive inspection, anti-counterfeit features, and contaminant detection that is invisible under visible light

The UV365/395 series covers the fluorescence case specifically: UV adhesive cure inspection and anti-counterfeit marks.

What to verify: whether your feature is a colour difference, a texture difference, or a fluorescence response. A white light on a UV-adhesive inspection produces nothing; the feature is only visible under excitation.

3. Uniformity across the field — not at the centre

Uniformity quoted without a field position is not usable. A light can be even at the centre and fall off 30% at the corners, which is exactly where a dimensional measurement is least reliable.

What to verify: the uniformity figure and the area over which it is specified, then compare that area against your field of view. A light specified uniform over 50 mm will not deliver that uniformity across a 200 mm field.

4. Size against working distance and field

The light must cover the field at the distance the fixture allows. Two constraints interact:

- Coverage — the emitting area must exceed the field of view at the mounting distance. - Clearance — the light must physically fit between the camera lens and the part.

The telecentric backlight series publishes its size range explicitly: VP-GLTD 30–265 mm and VP-GLTDL 35–265 mm. Match that against the outer dimension being measured — a backlight smaller than the part cannot produce a silhouette of it.

What to verify: the part's largest dimension against the light's specified size range, and the available mounting distance.

5. Control interface and intensity stability

Two things hide here.

Control: the light needs a controller with the right interface — constant current, triggerable strobe, or analogue dimming. A strobe synchronised to the camera exposure reduces motion blur and heat; continuous operation does not.

Stability: LED output drifts with temperature and with age. For a dimensional station, that drift changes edge position and therefore the measurement.

What to verify: whether the controller supports triggering from your camera, and whether the application can tolerate the warm-up drift.

Real application scenarios

A scratch inspection that lost its defect after a lighting change. The original was a bar light at a grazing angle; the replacement was a direct ring light. The scratch is still there, but grazing illumination is what raised it into contrast. The fix is geometry, not brightness.

An outer diameter measurement that drifts. A standard backlight produces a silhouette with edge blur that varies with part position. A telecentric parallel backlight removes the divergence. Check the size range against the part diameter.

A UV adhesive inspection that shows nothing. The adhesive is only visible when excited. Moving to a UV365/395 source makes the feature appear; no amount of white-light brightness will.

What this checklist does not tell you

- It does not state that any POMEAS light replaces any third-party light. That requires both datasheets and a test on the actual part. - Numerical uniformity, intensity and lifetime figures are per-model and belong on the product pages. Nothing is estimated here. - Third-party specifications are not reproduced. Nothing in this article states a specification for Moritex or any other lighting brand.

---

FAQ

Why did my defect disappear after I changed the light?

Most often the geometry changed. Scratches, embossing and texture need grazing illumination; a direct ring light flattens them out. Replicate the original geometry before adjusting brightness.

When do I need a telecentric backlight?

When measuring outer diameter or a profile and the part position varies. A standard divergent backlight produces edge blur that changes with position; a telecentric parallel backlight does not.

Is white light always the safe choice?

No. If the feature is a fluorescence response — UV adhesive, anti-counterfeit marks — white light will not excite it. If the feature is a fine texture on metal, blue often gives more contrast.

Does brightness matter more than uniformity?

Uniformity. A bright but uneven light produces measurements that vary across the field. Verify uniformity over the area you actually image.

Do I need a strobe controller?

For parts that move during exposure, or where heat and LED drift are concerns, yes. A strobe synchronised to the exposure freezes motion and reduces duty cycle.

How large should the light be?

Larger than the field of view at the mounting distance. For backlight silhouettes, larger than the part's outer dimension — check the published size range against it.

---

Next Step

Send the part material, the feature you need to reveal, the field of view and the available mounting distance. POMEAS will recommend the lighting geometry and wavelength from published specifications.