Machine visioning is a process whereby computer vision is applied to the process of development, production and manufacturing. Incorporating cameras, lighting systems, and image processing programs, machine visioning systems are capable of performing the inspection of items that would normally be performed by human beings. The reason for the explosive growth in machine visioning technology has been the ability of this technology to perform such inspection applications far more quickly and reliably than human beings as well as for long periods of time without any increase in error rates.

Machine visioning although effective is not an intelligent technology, meaning it cannot perform decision making and judgmental tasks the way a human operator can. However, machine vision systems do not get tired, they can repeat the same process over and over for long periods with little to no change in accuracy, and depending upon the parameters and capabilities of the software being used can read numbers and coding, detect flaws and defects in finished materials and count items as they are processed. Just like with humans however, the accuracy and effectiveness of machine vision can be acutely affected by the quality of the illumination being provided. As with humans, poor illumination can result in increased errors and inaccuracies, and thus providing an effective light source is critical to ensuring the effectiveness of machine visioning.

One of the most common types of illumination used in machine visioning involved the use of a halogen bulb based light source. This type of illumination proved effective and for many years has been used in a number of visioning applications with great success. The problem with halogen light sources however has been the need to include fiber optic light delivery systems for accurate beam placement, the relatively short 1500 hour lifespan of a halogen bulb, and the significant color shifting that takes place over the life of the halogen bulb. Beam placement is critical to accurate visioning operation, and maintaining intensity and light levels plays a large role in how efficiently a machine visioning system operates. The short lifespan of the halogen bulb means frequent replacements are necessary, which in turn lowers productivity, increases maintenance costs, and increases downtime leading to production losses. The significant color shifting of halogen bulbs also affects accuracy and reliability as it means that as the quality of the light produced by a halogen degrades, so too does the accuracy and reliability of the visioning system.

New lighting technology in the form of LED machine vision lighting has proven to be an affective way to address these shortcomings and led to a significant improvement in machine visioning effectiveness and reliability. Machine vision systems lack the sensitivity of the human eye and as a result require lighting that can enhance and highlight the features of items being processed. It is not enough to simply illuminate an object the same way we would for a human eye to see it accurately, illumination must take into account the reduced sensitivity of the cameras and sensors used in machine visioning. As noted earlier, changes in intensity and light quality can have a significant effect on how objects are perceived, and this effect is increased in the case of machine vision systems due to their lower sensitivity. If lighting quality and levels can be maintained at higher levels for longer periods, then the effectiveness of the visioning systems will be improved as well, and this is precisely the kind of performance LEDs are capable of providing.

Past machine vision systems relied on a halogen bulb light source and a fiber-optic delivery system to provide well positioned illumination where it was needed. Drawbacks to this setup included short bulb life and degradation of light quality. Current vision systems are incorporating LEDs which provide huge improvements in these areas, with operating lives surpassing 50,000 hours and color shifting and light intensity changes taking place at a much slower rate and over a much longer period of time. Red LEDs are currently the most popular color used in machine visioning systems, which is beneficial in that red LEDs tend to be the longest lived at 100,000 hours or so and the least expensive to produce.

LEDs can also be adapted over a wide range of color needs, with white, blue, and even ultraviolet and infrared LEDs being available for specific visioning applications. The benefit here is twofold in that not only is reliability improved with the use of LEDs, but there is no need for filters or lenses to alter the color output of the light source, which in turn simplifies assembly of machine vision systems and helps reduce costs. In some applications, such as those requiring backlighting of an object, getting enough light onto the target area meant trying to increase the light provided by the fiber-optic guides, which was difficult and cumbersome due to increased heat and issues with light distribution efficiency. With LEDs, increasing intensity and output is more simply a matter of using more powerful LEDs or a larger number of them in an array. Heat control with LEDs is typically built into the LED emitter assembly, and intensity can be increased greatly with a much lower corresponding increase in energy use which in turn also means less heat.

Although LEDs are typically a more expensive light source compared to standard halogen bulbs, machine visioning systems require a light source and fiber-optic array assembly which bumps the cost of the light source up considerably. An LED array taking the place of this bulb and fiber-optic array is similar in costs, which in turn means in terms of operating and maintenance costs versus operational life the LED array is actually a cheaper alternative as well as better performing. Additional benefits include the ability to control LED intensity and strobing via the use of an external controller, which adds greater flexibility and versatility.

Currently LEDs cannot fulfill all the needs of machine visioning systems, but are close to achieving this universal role. With LED development continuing at such a rapid pace and increasingly more powerful LEDs with higher color rendering capabilities become available, it is only a matter of time before LEDs become the dominant form of lighting technology in the field of machine visioning applications.