As the field of machine vision grows, it becomes increasingly important to understand what makes a machine vision system work well. The point of a machine vision system is to provide an image that makes it easy for a software package to do what needs to be done with said image. As such, it is important to have all of the parts of the machine vision system work in congruence. One of the most notable, and yet most often forgotten, parts of an imaging system is the lighting, or illumination. The camera and lens play an important part in the imaging system but these are more akin to the foundations of a building. In order to create a building worth mentioning, one needs to build on top of the foundations, this being illumination in an imaging system. Not unlike in construction, illumination is often what takes the most time when considering an imaging application. This is because lighting is so application specific. Specific camera and lens combinations can sometimes be used for different applications but very rarely does illumination geometry not vary between different applications.
When considering illumination, the object or field to be imaged is really what matters. There are so many different considerations but the important ones are as follows: transmissivity and surface texture. Transmissivity can be broken down into whether an object transmits light, absorbs light, or reflects light. Reflectivity can then can be broken down according to surface texture into specular reflections, or more mirror-like reflections, and diffuse reflection, or reflection more similar to how a piece of paper reflects. Knowing these properties makes it very simple to choose an illumination system that will maximize the efficiency of the imaging system.