Why Use Optical Filters in Machine Vision?
When developing a machine vision application, considerable attention is paid to the camera, lens and image processing software. However, reliable inspection starts with a good camera image. Ambient light, reflections and unwanted wavelengths can make product features difficult to distinguish. An optical filter can improve contrast before the image is processed by the software. Our technical selection guide for optical filters explains which filter is suitable for your application.
The following three applications show how optical filtering can improve industrial machine vision inspections.
Surface Cleanliness Inspection Using UV Fluorescence
When inspecting surface cleanliness, small contaminants such as dust particles and other residues need to be detected. Under conventional illumination, these can be difficult to distinguish from the surface.
Certain contaminants fluoresce when exposed to UV illumination. They absorb UV light and emit light at longer wavelengths, often within the visible spectrum.
With a suitable longpass filter, reflected UV light and unwanted stray light can be suppressed while the fluorescence emitted by the contamination reaches the camera. This improves contrast and makes it possible to inspect the presence and distribution of fluorescent contamination on the surface.
Not all contaminants fluoresce. The method should therefore be tested using the actual contamination and surface material.
Robot Picking Under Changing Ambient Lighting Conditions
In robot picking applications, the vision system must reliably detect products and determine their positions. On a production line, incoming daylight and lighting from nearby machines can affect the camera image.
By illuminating the products with LEDs at a specific wavelength and mounting a matching bandpass filter on the lens, a large portion of the ambient light can be suppressed.
This makes image contrast less dependent on changing lighting conditions and can improve the reliability of product segmentation and pick position determination.
Moving products often require short exposure times to avoid motion blur. The bandwidth and transmission of the filter must therefore match the LED illumination to ensure that sufficient light reaches the camera. Our range of bandpass filters for machine vision provides options for different illumination wavelengths.
Reading Data Matrix Codes and Barcodes Through Transparent Film
When reading Data Matrix codes (DMC) and barcodes through transparent packaging film, reflections from the illumination can overexpose parts of the code. This reduces contrast and can make the code more difficult to read.
One solution is cross-polarisation. A polarising film is placed in front of the illumination, while a second polarising filter is mounted on the camera lens.
By adjusting the polarisation directions relative to each other, specular reflections from the film can be significantly reduced. This improves code visibility and can increase reading reliability.
Not all transparent films respond in the same way to polarised light. Some plastic films alter the polarisation state, which can reduce the effectiveness of cross-polarisation. Test the setup with the actual packaging material and suitable polarising filters for industrial cameras.
Optimise the Camera Image Before Image Processing
In machine vision, it is often easier to solve an imaging problem optically than to compensate for it later through additional image processing.
A correctly selected filter can suppress unwanted light and make relevant product features easier to distinguish. This reduces the need for software-based corrections during segmentation, position measurement and code reading.
However, an optical filter cannot replace a properly designed illumination setup. Illumination geometry, material properties and the spectral sensitivity of the camera remain critical to the result.
Which Optical Filter Do You Need?
The choice between shortpass, longpass, bandpass and polarising filters depends on which wavelengths or reflections need to be transmitted or suppressed. Our optical filter selection guide explains the technical differences, provides application examples and helps you select the correct filter thread size for your lens.
If you already know which filter you need, explore our range of optical filters.
At Machine Vision Shop, we support engineers in selecting and testing optical components for industrial machine vision applications.
