Optical Filters for Machine Vision: Technical Selection Guide

An optical filter can suppress unwanted light and improve contrast in a machine vision inspection. The appropriate filter depends on the illumination, the spectral sensitivity of the camera and the properties of the material being inspected.

Below, we explain four important filter types with examples from industrial machine vision. Once you have selected the appropriate filter type, you can find the available filters by the lens filter thread size.

Which type of optical filter do you need?

Shortpass Filter

A shortpass filter transmits wavelengths below a specified cut-off wavelength and blocks longer wavelengths. One common application is removing unwanted near-infrared (NIR) light from the camera image.

Application example: Colour inspection

Many industrial image sensors are also sensitive to near-infrared light. In colour cameras without an integrated IR-cut filter, this can affect colour reproduction. A shortpass filter that transmits visible light while blocking NIR prevents unwanted infrared light from interfering with colour measurements. Uniform transmission across the visible spectrum is important.

Longpass Filter

A longpass filter transmits wavelengths above a specified cut-on wavelength and blocks shorter wavelengths. This is particularly useful for fluorescence inspection and applications using infrared illumination.

Application example: Reading codes printed with UV fluorescent ink

When reading codes printed with UV fluorescent ink, the ink is excited by UV illumination. It absorbs the UV light and emits light at a longer wavelength, often within the visible spectrum.

A suitable longpass filter blocks the UV light and unwanted violet or blue stray light from the illumination while transmitting the fluorescence emitted by the code. This increases contrast between the code and the background, making the code easier to read.

The cut-on wavelength of the filter must be below the relevant emission range of the fluorescent ink.

Application example: Licence plate recognition with IR illumination

In licence plate recognition (LPR/ANPR), a longpass filter can be combined with infrared illumination. The retroreflective properties of many licence plates cause IR light to be reflected strongly back towards the camera.

The longpass filter suppresses visible ambient light. Combined with the correct illumination geometry, this produces an image in which the licence plate appears bright against a much darker background.

This simplifies licence plate segmentation and can also benefit AI-based LPR systems. However, some AI models rely on additional vehicle information to locate the licence plate.

Bandpass Filter

A bandpass filter transmits a limited wavelength range while suppressing light outside this range. This allows the camera to be matched to the wavelength of the LED or laser illumination.

Application example: Inspection with green LED illumination

When using green LED illumination around 525 nm, a suitable bandpass filter can block a large portion of ambient light. This makes the inspection less sensitive to changing lighting conditions.

Check that the centre wavelength (CWL) and bandwidth (FWHM) match the emission spectrum of the LED. A passband that is too narrow may also block part of the desired light, requiring a longer exposure time or higher LED intensity.

Polarising Filter

A polarising filter selects light based on its polarisation direction. In machine vision, it is mainly used to reduce unwanted specular reflections.

Application example: Inspection of glossy plastic packaging

Reflections from glossy packaging can overexpose parts of printed text or barcodes. By placing polarising film in front of the illumination and mounting a polarising filter on the lens, these reflections can be significantly reduced through cross-polarisation.

The result depends on the material and illumination geometry. This method is not always equally effective on metallic surfaces. In some cases, adjusting the illumination angle or using diffuse lighting produces better results.

What should you consider when selecting an optical filter?

When selecting an optical filter, the following properties are particularly important:

  • Wavelength: the transmission range must match the illumination and the spectral sensitivity of the camera.

  • Transmission and bandwidth: determine how much desired light is transmitted and how much unwanted light is blocked.

  • Mounting: check the filter diameter, thread pitch and clear aperture to avoid vignetting.

A filter cannot compensate for incorrect illumination geometry. First evaluate how the light reaches the product and which reflections or wavelengths actually interfere with the inspection.

Select an Optical Filter by Thread Size

Many industrial lenses have a filter thread at the front. The filter size is specified by the nominal thread diameter, for example M30.5.

Select the filter thread size of your lens below to view the available optical filters and wavelengths.

Available Filter Thread Sizes

  • M22.5 optical filters

  • M25 optical filters

  • M25.5 optical filters

  • M27 optical filters

  • M30.5 optical filters

  • M34 optical filters

  • M35.5 optical filters

  • M37 optical filters

  • M37.5 optical filters

  • M40.5 optical filters

  • M43 optical filters

  • M46 optical filters

  • M49 optical filters

  • M52 optical filters

  • M58 optical filters

Note: always check the thread pitch as well as the diameter. M30.5 specifies only the nominal thread diameter. M30.5 x 0.5 also specifies a thread pitch of 0.5 mm.

Technical Advice on Optical Filters

The right combination of illumination and optical filtering can significantly simplify image processing. At Machine Vision Shop, we support engineers in selecting and testing optical components for industrial inspection applications.

Explore our optical filters or contact us to discuss the filter selection for your application.