The central role of lighting in machine vision
Poor lighting is the most common cause of poor machine vision performance. Even sophisticated cameras and software can’t make up for inadequate lighting.
Good lighting for machine vision maximizes contrast on the features of interest and minimizes it everywhere else. It also needs to be consistent; the lighting setup should make sure normal variations in parts or their arrangement don’t affect the contrast in the image.
Depending on the part’s texture (such as matte, reflective, or refractive) and its shape, different lighting choices will change which portions of the image will be brighter than others. Poorly chosen lighting might create glare when light reflects off a smooth part directly into the camera. A different position might occlude an important spot or leave it in a shadow.
Good lighting for machine vision maximizes contrast on the features of interest and minimizes it everywhere else.
Some cameras come with integrated lights, which provide uniform illumination that’s important when working with matte objects, among other uses. They’re easy to deploy and work well in many situations.
Other conditions can require a separate lighting setup, which can involve experimentation with different lighting angles and types to get the most effective image. There’s no one best lighting setup for all machine vision applications. Instead, lighting optimization depends on the application, specific requirements, and the desired results.
Types of machine vision lighting
Backlighting – A flat, diffuse light behind the target forms a black silhouette with maximum contrast and clear edges.
Full/partial bright field lighting – Lighting an object from above creates strong contrast so flat parts reflect into the camera.
Dark field lighting – Shining a light from the side at a low angle (like a rising or setting sun) helps show edges, height variations, texture, and contours. Background areas reflect light away from the camera and appear dark, while the camera receives light from the features of interest.
Diffuse lighting – Even lighting from multiple directions minimizes confusing reflections from a curved or shiny surface. Also called cloudy-day lighting, diffuse lights can be domed, axial, or flat, and help create an even background, making it easier to detect features of interest.
Multicolor/multispectral lighting – Using infrared (IR) lighting can make it easier to find small surface defects within visually busy images, see differences in material, and work in environments with challenging visual conditions.
Filters
Filters can keep unnecessary ambient light from entering a camera or increase the contrast of a part or other object.
- Color filters only let red, green, or blue light through. Depending on the objects being examined, one color may provide higher contrast.
- IR and UV filters only let through light outside the visible spectrum and can often show contrast differences that don’t show up in visible light.
- Polarizing filters can eliminate glare from light being reflected into a camera.
Color filters can heighten contrast for certain colored objects.
Polarizing filters can reduce glare.