Midwest Optical Filters: Machine vision experts know it’s a necessity, not an accessory
In industrial imaging, achieving a perfect shot isn't just about the camera or the lens; it’s about controlling the light that reaches the sensor. Midwest Optical Filters have established themselves as a leading filter solution in machine vision for control and repeatability.
This is the 5th and final part of our series on Midwest Optical filters.
In parts 1 and 2 series we described how optical filters can be used to advantage when using monochrome cameras
- Mimic specific LED behaviour for testing purposes
- Block unwanted light wavelengths to increase contrast
- Reduce lens chromatic aberration and increase resolution
- Reduce reflected glare to assist with brightness control
- Increase contrast by using non-visible near infrared wavelengths
- Enable UV or Fluorescence imaging
- Provide a mechanical barrier to protect the front of a lens
In parts 3 and 4 of this series we discussed how filters can be used with colour cameras to assist in improving results
- Produce more natural colour images
- Reduce glare
- Balance colour from light sources to achieve required colour temperatures
- Adjust image brightness
- Difference indexing calculations (e.g. NDVI)
- Enable day and night imaging with a single camera
Bandpass filters
Bandpass filters are often the simplest way to see a dramatic improvement in image quality. They are designed to pass a single band of wavelengths. They can feature Gaussian transmission curves or more square curves with sharp cut-offs at the edges. They can be very narrow and broader wavelength passing filters and this is denoted by the FWHM (Full Width Half Max) number. Bandpass filters are perhaps the most commonly used optical filter type in machine vision.
Here is an example of how a broad blue bandpass filter can be used to improve contrast for this specific application.
Multi Bandpass filters
Multi Bandpass filters transmit two or more specific bands of wavelengths and are perhaps most commonly used in combination with colour cameras and NIR lighting to enable day time colour images and night time NIR images. Intelligent traffic applications benefit greatly by using this technique.
Multi Bandpass Filters are also great for outdoor security or agricultural inspections. You can achieve natural colour rendition during the day and clear near-infrared imaging at night without needing to swap hardware, providing a 24-hour imaging solution.
Longpass filters
Longpass filters block shorter wavelengths and allow longer wavelengths to pass. They are an economical and effective tool to eliminate wavelengths that are present and sometimes necessary but interfere with the imaging inspection.
An example of this is the use of longpass filers in fluorescence applications. By blocking the shorter wavelength and typically very bright light targeted at the material which causes it to fluoresce, and by passing the longer weaker fluorescence wavelengths emitted by the material, these filters provide much higher contrast imaging than would otherwise be possible.
Shortpass/NIR cut filters
Shortpass filters block longer wavelengths while passing shorter wavelengths. They are almost always part of single sensor colour machine vision cameras to ensure true colour imaging. These cameras use bayer mosaic colour sensors where every pixel is overlaid with an R, G or B filter at the sensor level. These filters pass the bands of light that correspond to each of these primary colours, however they also often pass NIR wavelengths which if not blocked upset the true colour balance.
Polarising filters
Polarising filters are the industry standard for reducing glare and reflections from shiny or specular surfaces.
Light that hits a specular surface reflects partially polarised light and in the image above presents as bright while reflections on the water’s surface. Polarisers when cross-aligned to the polarisation angle of this reflected light block it to remove the unwanted reflections.
Polarising Filters can also provide advantages with machine vision applications for example when you are inspecting products wrapped in plastic or checking for stress marks in transparent glass.
Neutral Density (ND) filters
ND filters partially block all wavelengths of light across the image without affecting colour. ND filers vary in their density and subsequently the percentage of light they passed. The correct ND filter can be chosen for the task. They are for example another way of controlling brightness in an image without adjusting the lens aperture. The lens aperture may be set wide open in order to reduce the depth of field and so there may be too much light, an ND filter serves to reduce the light ad so avoid saturation in the image.
Neutral Density Filters can provides much greater control over your exposure settings in bright environments.
Protective filters
Available in high-quality glass or optical-grade acrylic, these filters act as a physical barrier for your expensive lenses to avoid them getting damaged.
Protective Filters also offer advantages in harsh industrial or food-grade environments (FDA/EMA applications), where an acrylic protective filter can prevent glass shards from contaminating a production line if an impact occurs, saving you from both equipment damage and costly recalls.
Light balancing filters
These filters adjust the "temperature" of your light to achieve a more task appropriate colour rendering, particularly when using cool white LEDs.
Light Balancing Filters can block the harsh blue peaks often found in LED lighting, these filters make the inspected object look more natural. This is essential for human-in-the-loop systems where operators need to see a realistic representation of the product.
Precision control for machine vision
Midwest Optical Filters deliver the repeatability and spectral accuracy required for high-stakes industrial inspection. Contact our engineering team to select the right interference or neutral density filters to optimise your imaging consistency.