Basic Illumination Techniques for Machine Vision

Even with the most powerful technology, machine vision systems won’t work without the right lighting. A system’s lighting needs are highly variable, depending on the specific application and the operating environment. Engineers must be knowledgeable and flexible enough to optimize lighting based on the desired outcome and their knowledge of optics.

Fundamental illumination techniques include:

Back Lighting

By creating dark silhouettes against a bright background, back lighting generates immediate contrast. Back lighting can be used for a variety of common inspection tasks: For example, part placement and orientation, detecting the presence or absence of gaps, or measuring simple objects. If precise edge detection is a requirement, the engineer should consider a monochrome red, green, or blue light that features light control polarization.

Diffuse Lighting – “Full Bright Field”

When samples are shiny or feature mixed reflectivity, diffuse lighting is a method of delivering illumination that is both even and multi-directional. There are multiple effective implementations of diffuse lighting. Diffuse dome lighting is ideal for lighting specular, curved surfaces. By contrast, on-axis lights are preferred when working with flat surfaces. Diffuse lights must be situated in close proximity to the sample object.

Directional – “Partial Bright Field”

This is the most common lighting technique, and can be adjusted for effectiveness in a wide range of operating environments. It is also familiar and intuitive – technically, even sunlight is directional lighting! This kind of light derives from a point source, allowing it to generate high contrast and enhance the clarity of surface topography. It should not be used with specular surfaces, since “hotspot” reflection can severely compromise the final image quality.

Dark Field

Dark field is the least-used and perhaps least-understood method of illumination. It uses a medium or low angle of light incidence that demands close proximity – for example, headlights rely on dark field illumination principles. Dark field lighting can be subdivided into circular and linear types. Linear dark field lighting requires the closest proximity. For that reason, it can be the most difficult lighting solution to design and use effectively.

No matter what machine vision technologies or applications an engineer works with, lighting remains fundamental. To learn more about how lighting affects machine vision system design, access the AIA Vision Online Beginning Lighting for Machine Vision Part 1 video. It’s available for free, without registration, for all current and aspiring machine vision professionals.