A report released by the American market research consulting firm Transparency Market Research indicates that the global machine vision technology market is expected to grow from $15.7 billion in 2014 (approximately 104 billion RMB) to $28.5 billion in 2021 (approximately 189 billion RMB), at a compound annual growth rate of 8.40%.
Machine vision systems are indispensable for achieving factory automation, including assembly positioning, quality inspection, product identification, and dimension measurement. The high precision requirements of high-speed production lines can no longer be met by the human eye.
How can we reduce costs? There are many imaginable spaces, and embedded machine vision systems will be a good choice, with the core of the system certainly relying on microprocessors.Since the 1990s, microprocessors and semiconductor technology have been spiraling upwards alongside machine vision technology: microprocessors and semiconductor technology are the birthplace of machine vision, with extensive applications of image technology in Europe and America, which gradually evolved into today’s machine vision technology. The widespread application of machine vision in Europe and America is also mainly reflected in the semiconductor and electronics industries. Now, microprocessor performance has become stronger, power consumption has greatly reduced, and sizes have become more compact, yet prices have not increased.
Take the Raspberry Pi 3 microprocessor launched in February this year as an example, equipped with a 64-bit 1.2GHz quad-core chip and 1GB of memory, its performance has improved by 50% compared to the Raspberry Pi 2. Yet the price remains at $35 (approximately 232 RMB), the same as the launch price of the Raspberry Pi Model B four years ago. Mass applications of high-performance, low-power, compact microprocessors in embedded vision systems can potentially halve the individual costs.
Integrating machine vision systems on general-purpose computers involves multiple technologies such as lighting, imaging, image digitization, image processing algorithms, and hardware/software, which places high demands on technical personnel. Using embedded machine vision systems, the configuration of hardware and software becomes flexible, and the development environment and programs are more universal. This facilitates mass production and line expansion, greatly enhancing production flexibility, and the general demand for machine vision technology will be responded to quickly by enterprises.
The emergence of global standards related to machine vision has accelerated the progress of embedded systems.In June last year, the China Machine Vision Industry Alliance (CMVU) joined the G3 standard, becoming the 15th member unit of G3 standards. Other members include: American Automation Imaging Association (AIA), European Machine Vision Association (EMVA), Japan Industrial Imaging Association (JIIA), and German Mechanical Engineering Industry Association (VDMA). The publication of the “Global Machine Vision Interface Standards” handbook and the signing of the “G3 Camera Standards” protocol will shorten development time, reduce investment costs, and accelerate product market entry.
In addition, to make machine vision adaptable to “Industry 4.0” and future factory production, the VDMA Machine Vision Association and the OPC Foundation have begun to compile the “OPC Unified Architecture Machine Vision Supporting Specification,” aiming to directly integrate machine vision systems into production control and IT systems to maximize efficiency.
Industry 4.0 is about connecting production technology and information technology, and machine vision is one of the most important foundational technologies providing information for Industry 4.0. Embedded systems will play an increasingly important role in future machine vision systems. They can achieve compact product designs, meet most image processing requirements, and have a higher degree of integration with factory processes compared to PC-based independent systems.
Author: CEC Shi Lincai
