Manufacturing automation today is implementing exciting new technologies. Advances in artificial intelligence are providing the tools necessary to improve product quality dramatically while reducing the costs of production. One such tool that is invading production lines is machine vision technology. Optical analysis techniques are impacting both the quality of finished goods and the throughput of manufacturing processes. Machine vision is a term used to describe the process of collecting and analyzing optical data about objects in ways that enable advanced automation.

One industry that is focused heavily on this area is autonomous vehicle operations. Driverless cars are making their debut on highways around the globe successfully avoiding collisions and obeying traffic laws without human intervention. This is only possible with advanced machine vision. Interestingly, the automotive manufacturing industry is also using these techniques in applications such as bead inspection on assembly lines. Visual data about the application of adhesives used to add trim pieces and surfaces in car interiors are collected and analyzed to reduce costs and improve quality. Two basic elements of machine vision are the optical portion and the algorithmic portion.

Optical Element of Machine Vision

Not only are lens improvements being made to examine smaller and smaller details but the sensors behind the lenses are getting more sophisticated as well. Primarily the size of sensors and the density of elements on them are making higher resolutions possible. The ability to collect light frequency spectrums outside the visual region such as infrared and ultraviolet are adding valuable data from these regions of the spectrum.

Algorithmic Element of Machine Vision

The most advanced visual data is only useful when it can be interpreted in ways that will improve processes. Graphics processors have steadily gained more power as the information age has progressed. With these advanced hardware systems, programming techniques in areas such as deep learning and neural networks push computing capabilities ever closer to human cognitive abilities. Applying these techniques to machine vision has steadily endowed computers with more human-like perception capabilities. One area in which this is demonstrated is in the Imagenet competition. Here, machines are pitted against one another to see which can best identify specific objects or visual data.

Industries are leveraging the advances in machine learning and machine vision technologies to improve their operations. As technologies continue to evolve, ever more exciting capabilities are sure to become a reality.