Embedded technologies assist machine vision power, flexibility

Control Engineering China: Using embedded technologies in machine vision systems makes the configuration of software and hardware more flexible, the development environment and programming more universal, and greatly improves production flexibility. Standards are helping.

By Stone Shi February 10, 2017

Through the accumulated development of advanced embedded technologies over many years, machine vision is in a golden development period; wider use of embedded technologies continues to ease programming and increase flexibility. Machine vision costs are reduced, performance is improved, and applications have increased. Just like the reform brought by the popularization of wireless networks, machine vision software and hardware products will become standard configuration in manufacturing industries.

The U.S. survey and consulting firm Transparency Market Research expects the global machine vision technology market to increase to $28.5 billion (approximately RMB 189 billion) in 2021 up from $15.7 billion (approximately RMB 104 billion) in 2014, with a compound average annual growth rate of 8.4%. The machine vision system is indispensable for factory automation, assembly positioning, quality inspection, product identification, dimensional measurement, and other applications. The human eye cannot meet the requirements of high precision on a high-speed production line. 

Machine vision cost, features

How have machine vision costs continue to decrease as capabilities increase? The embedded machine vision system, a good choice for an increasing number of applications, needs a microprocessor. Since the 1990s, microprocessor and semiconductor technologies have spirally improved along with machine vision technology. Microprocessor and semiconductor technologies are the birthplace of machine vision. In Europe and America, a large number of image technologies are applied. Later, they have slowly evolved into the machine vision technology in use today. Machine vision is widely applied in Europe and in the U.S., especially in the semiconductor and electronics industries. At present, the performance of microprocessors becomes stronger, power consumption is greatly reduced, dimensions are more compact, and prices generally have not increased.

Take the microprocessor Raspberry Pi 3, which became available in February 2016, for example. Equipped with 64-bit 1.2GHz four-core chip and 1GB RAM, its performance improved by 50% compared to Raspberry Pi 2, while its price is still $35 (approximately RMB 232), which is the same as the pricing of Raspberry Pi Model B when first released four years ago. With the application of a great number of compact microprocessors with high performance and low power consumption for use in embedded vision systems, per item costs might be expected to reduce by half once more. 

Vision integration challenges

When the machine vision system is integrated on a general-purpose computer, very high requirements are presented for technical personnel, since integration involves multiple technologies, including lighting, imaging, image digitization, image processing algorithm, software, hardware and more. Using embedded machine vision systems makes configuration of software and hardware more flexible, and the development environment and programming become more universal. Embedded machine vision eases mass production and line expansions, greatly improves production flexibility, and enables faster responses to the general demand of enterprise systems connected to the vision-enabled plant-floor systems.

Helpful machine vision standards

The successive release of relevant global standards on machine vision speeds the progress of being brought into the embedded system. In June last year, China Machine Vision Industry Alliance (CMVU) joined G3 standard, making CMVU the 15th unit member of G3 standard. Other members include AIA, European Machine Vision Association (EMVA), Japan Industrial Imaging Association (JIIA), and VDMA  (Verband Deutscher Maschinen- und Anlagenbau, Mechanical Engineering Industry Association).

Development of the "Global Machine Vision Interface Standards" brochure and the signing of the "G3 Standardization Initiative" agreement will shorten development periods, reduce investment costs, and speed up product availability in the market.

To adapt machine vision to Industrie 4.0 and future factory production, VDMA Machine Vision and OPC Foundation have begun to formulate the "OPC Supporting Standard of Machine Vision with Uniform Framework" for the purpose of integrating the machine vision system directly into the production control and IT system, and maximizing effectiveness.

Industrie 4.0 is related to the connection of production technology and information technology, while machine vision is one of the most important supporting technologies to provide information for Industrie 4.0. The embedded system will play more roles in the future machine vision systems, including more compact product designs, and meeting wider image processing applications. Besides, compared to a PC-based independent system, it is more easily integrated with factory processes.

Stone Shi is executive editor-in-chief, Control Engineering China; Edited by Mark T. Hoske, content manager, CFE Media, Control Engineering, mhoske@cfemedia.com.


Key concepts

  • Embedded system technologies help machine vision.
  • Standards organizations for machine vision are cooperating.
  • Embedded vision technologies further Industrie 4.0 connectivity.

Consider this

What application could move faster with higher quality by using embedded machine vision technologies?

ONLINE extra

Control Engineering has a page for machine vision and sensors

See more Control Engineering international coverage.

See Control Engineering China coverage.

See related articles on machine vision below.

Author Bio: Executive editor-in-chief, Control Engineering China