Smart phones, tablets, hand-held devices and smart cameras — the use of mobile devices is becoming more widespread even in the industrial environment. To what extent are these devices significant for machine vision? And how can we make the best possible use of embedded technologies?

There have been times when machine vision was not a classic application field for mobility applications. This was simply due to the fact that the computing power needed for this technology far exceeded the performance of the mobile devices available at the time. This situation has changed. Now there are smart phones and tablets, as well as smart cameras and embedded vision systems on the market that can be used to perform even complex machine vision tasks without problems. And the performance of mobile devices continues on a steep upward trend in terms of storage capacity, display technology, and processor performance. Although high-end smart phones and tablets come with high-resolution cameras, these cameras are usually not suitable for use in demanding machine vision applications. Smart cameras or embedded systems with powerful industrial cameras are therefore used in an industrial setting.

Mobile devices are penetrating industrial processes

In addition, the importance of mobile devices, which used to be more commonly found in the consumer sphere, is increasing rapidly in production environments on the way to Industry 4.0. Embedded vision systems are therefore becoming more and more significant. Integrating a large number of different embedded platforms, such as Raspberry Pi or the ARM processor architecture, into existing environments is a challenge.

Powerful machine vision software, which can run on these compact devices without problems, is required for the targeted use of these mobile helpers in an industrial setting. We offer practical technology for this purpose. It is ideally suitable for use in embedded systems that work with the popular ARM architecture. We therefore also provide our HALCON standard software for the Android operating system and thereby place complex applications, such as comprehensive machine vision scenarios, on mobile devices.

Snapshots of optimized maintenance processes

This paves the way for a great many flexible machine vision applications within the scope of industrial processes. For example, smart phones, phablets, and tablets can be used to take quick snapshots when repairing and maintaining machines. If a particular component must be replaced in a control cabinet, the service technician simply points his mobile device toward the cabinet. The machine vision software detects the component to be replaced and reliably identifies it. The technology also provides assistance when preparing a so-called multimedia manual. Employees can detect different components with the aid of their mobile devices and thereby access instructions for certain maintenance steps or additional information on-line. These are only a few examples of how mobile machine vision processes can speed up and optimize industrial processes.

Embedded vision for smart cameras and mobile vision sensors

We should not neglect to mention the many machine vision systems which still exist today in industry and which are used to continuously inspect products and monitor production processes. These “traditional” machine vision systems have up to now been permanently connected to industrial PCs and use one or multiple cameras. With the aid of innovative embedded technologies, machine vision software can also be used for this segment in smart cameras and mobile vision sensors. They can be flexibly exchanged and therefore respond more quickly to changed product requirements. Embedded technologies therefore significantly increase the possible uses of machine vision applications.

Dr. Olaf Munkelt is co-founder of MVTec Software GmbH and, since its founding in 1996, also one of the company’s managing directors. He has served as chairman of the board of directors of the machine vision group of the German Engineering Federation (VDMA) since 2009. Before starting MVTec, Olaf was a research assistant at the Technische Universität München (TUM) and graduated with a PhD in computer science in 1994. After that he was acting as co-head of the research group for cognitive systems at the Bavarian Research Center for Knowledge-based Systems (FORWISS) for three years.