The Window in Front of the Lens

Generic engineering material, applicable to any module from any supplier.

Additional workflow context is available further details.

A protective window is an optical component, and treating it as a piece of plastic in a housing is how a good imaging design acquires reflections nobody can remove afterwards.

What a window does

Reflects. Every surface reflects a few per cent, and a window has two — so light from the scene bounces between them and back off the lens, producing a faint copy of the image offset from the real one.

Which is ghosting, and it is most visible against dark regions and around bright sources.

Further background is available from Optics.org.

Absorbs, slightly, which matters only where light is already scarce.

And distorts, where the material is not flat or not uniform in thickness — an injection-moulded window is not an optical flat unless somebody specified it as one.

The angle that removes the ghost

Tilt the window a few degrees relative to the optical axis.

The reflection then travels away from the lens rather than back into it, which removes the ghost at the cost of nothing.

A few degrees is enough in most geometries, and the direction matters: tilt away from any bright source in the scene rather than toward it.

This is the single most effective thing on this page and it costs a change to a drawing.

Coatings

An anti-reflection coating reduces the reflection at each surface substantially.

Worth specifying where the window is unavoidably perpendicular, or where a bright source is unavoidably in frame.

And it is a maintenance consideration, since a coated surface is more easily damaged by wiping than an uncoated one — which argues for the tilt where the environment is dirty.

Material

Glass is harder, more scratch resistant, optically better and breaks.

Plastic survives impact, scratches easily, and varies more in optical quality — and some plastics attenuate near infrared, which matters where an infrared illuminator is part of the design.

Check the transmission at your working wavelength rather than assuming a clear material is clear at it.

And check chemical compatibility with whatever will clean it, which in washdown environments is the deciding factor.

The gap behind it

Dust settles on the inside surface and cannot be reached.

A gap also collects condensation on the coldest surface, which is usually the window.

Which is why the sealed optical chamber described elsewhere is assembled dry with a desiccant or vented — the gap is where the problem lives.

And a narrow gap is better than a wide one for both, since there is less air and less surface.

Where the window is behind glass already

Cameras looking through a vehicle windscreen, a display case, or a machine guard.

Which adds a second window nobody specified, with its own reflections, its own tint and its own cleanliness state.

And frequently a curved one, which distorts in a way a flat plate does not.

Two consequences. A shroud between the lens and the outer glass removes most of the reflected interior — the standard remedy on any camera behind a windscreen.

And the outer surface is somebody else's to clean, which belongs in the maintenance conversation rather than being discovered.

Aperture size in the housing

The opening in front of the lens must clear the whole field of view.

Which sounds obvious and is a common error, particularly with wide-angle modules: an aperture sized by eye vignettes the corners, producing dark edges that look like a lens defect.

Calculate it from the field angle and the distance between the lens and the opening — and add margin for the mounting tolerance, since the module may not sit exactly where the drawing says.

A test image of a uniform surface reveals it immediately, and it is worth doing on the first mechanical prototype rather than the first production run.

Internal reflections

A shiny housing interior reflects light onto the sensor.

Which raises the level in dark regions, exactly as contamination does, and from a cause nobody looks for.

Matte black finish inside the optical chamber, and where possible a baffle between the lens and any bright interior surface.

It costs a paint specification and it is the difference between a design that looks right and one that measures right.

Specifying it

Four things on the drawing.

Material and grade, with transmission at the working wavelength where infrared is involved.

Thickness and flatness tolerance, since an unspecified moulding is not an optical flat.

Tilt angle relative to the optical axis, which is the item that removes the ghost and the item most likely to be omitted.

And the finish of the aperture and chamber, matte black.

Four lines, and they are the difference between a window and a piece of plastic in a hole.

Testing for ghosting

Photograph a bright point source against a dark background.

A torch in a dark room is sufficient, positioned at several places in and just outside the frame.

A faint second image, displaced from the real one, is the reflection this page describes — and its position tells you which surface is producing it.

Do it on the first mechanical prototype, since the remedy is a tilt angle and that is a drawing change rather than a rework.

In one line

Tilt it a few degrees, which costs a drawing change and removes the ghost entirely.

The short version

• A protective window is an optical component: it reflects at two surfaces, absorbs slightly, and distorts if it is not flat
• Reflections between window and lens produce ghosting, most visible against dark regions and near bright sources
• Tilting the window a few degrees sends the reflection away from the lens and costs nothing but a drawing change
• An anti-reflection coating helps where a perpendicular window is unavoidable, and is more easily damaged by cleaning
• Glass is harder and optically better; plastic survives impact and some grades attenuate near infrared
• The gap behind the window collects unreachable dust and condensation, which is why it is sealed dry or vented