1. Machine Vision: The key to 5G industrial era and intelligent manufacturing
Machine vision is a device that automatically receives and processes images of a real object through optical devices and non-contact sensors to obtain the required information or for controlling robot motion.
Industrial vision has extremely strong machine characteristics, such as extremely high speed, accuracy, and repeatability, which become important parameters for industrial machine vision products.
At the same time, industrial machine vision has the ability to collect large amounts of field data, and data is an important source for driving rapid industry iterations and obtaining information, which will become a huge possibility for the future of industrial machine vision.
5G enables the long-term development of machine vision, opening an important era of industrial connectivity
Industrial machine vision is an important part of the industrial production environment, and 5G provides the network foundation for industrial machine vision to scale up, ultimately facilitating the building of the entire industrial Internet.
Machine vision has reached 6.5 billion in 2019, and will develop at a high rate in the future
According to Deloitte's "China Artificial Intelligence Industry White Paper", the global artificial intelligence market has reached $680 billion, of which the size of the Chinese market is $71 billion. According to Sadie Consulting data, by 2020, China's intelligent manufacturing equipment market is expected to reach 2 trillion yuan.
GGII data show that China's machine vision market size in 2019 is 6.550 billion yuan (the data does not include computer vision market size), an increase of 21.77%. 2014-2019 compound growth rate of 28.36%. GGII forecasts that by 2023 China's machine vision market size will reach 15.56 billion yuan.
2. Upstream industry chain summary: the pace of hardware localization accelerated; the software still needs to be cultivated
Because of the early development of overseas markets, the upstream components industry is still basically occupied by overseas enterprises, Chinese enterprises after the development of recent years to catch up, and are expected to gradually break through from the relatively low technical barriers of light sources and industrial cameras. Grasp the trend of domestic substitution and occupy the market through cost-effective products.
The lens involves multi-disciplinary technology, complex production process, the high requirements for precision instruments, the high initial investment, the international market is still firmly held by the old manufacturers. Chinese companies are currently competing with international manufacturers with the advantage of high cost performance. With the development of the industry and the accumulation of technology and iteration, domestic manufacturers are expected to gain a foothold in the long run.Centralization (smart cameras) is one of the technology trends for upstream components
The combination of light source, lens, industrial camera, image processing software and other components for the traditional PC-based industrial vision system. With the development of embedded technology in recent years, smart cameras have gradually gained attention as relatively completed solutions. Smart cameras are highly integrated micro-compact machine vision systems, and the finished product has integrated image acquisition, image processing unit and software, and network communication within a single camera.
Smart cameras are low cost, versatile and easy to use, but the accuracy, speed and ability of complex computing are far less than traditional systems, so they are especially suitable for small and medium-sized private enterprises that are price sensitive.
3. Downstream applications continue to expand, industrial machine vision has great potential
Automotive intelligence to promote the application of machine vision. In the field of automotive manufacturing, the cost of electronic components in new energy and smart cars will reach more than half of the entire vehicle, a large number of sensors, communications (GPS, DSRC, 4G/5G), cameras, monitoring, detection, entertainment systems will be loaded on top of the car. The automotive industry chain's requirements for production accuracy and intelligence are increasing.
Electronic product iteration to promote the application of machine vision. With the successive production lines of high generation panels, flat panel display devices to intelligent, large size, thin, touchable, high resolution, flexible panels, self-luminous, high migration rate and low power consumption has reached a consensus in the industry.
Semiconductor industry progress and domestic substitution jointly promote the development of machine vision. In the field of semiconductor manufacturing, domestic IC technology breakthroughs in new process nodes and the landing of 5G technology bring a new round of boom in the domestic semiconductor industry. At the same time, the semiconductor manufacturers' motivation for equipment localization continues to rise, bringing opportunities for Chinese machine vision companies.
Horizontal expansion of emerging fields. With the progress of the industry and technological development, 2D vision to 3D vision, thus expanding more new areas, such as robot arm guidance, AGV navigation, etc.. Machine vision will gradually cut into the past uninvolved areas, such as logistics, medical, security, agriculture and other non-industrial scenarios; at the same time, the basic technology will deepen the penetration of machine vision in the automotive, consumer electronics, semiconductor and other industrial fields.
4. Industrial machine vision future trends forecast
The industry of industrial machine vision is driven by a combination of technology, downstream industries, and social structure.
Machine vision is produced in line with the development of industrial automation, born with high precision, high speed characteristics of the machine, in special applications such as panel production, semiconductor, machine vision has an irreplaceable role, not just a simple "cost reduction and efficiency".
Downstream industries continue to iterate vertically (such as the intelligence of the automobile, panel field screen iteration), horizontal expansion (such as semiconductors, new energy, industrial robots, etc.), the demand for machine vision more and more huge.
In the past, the domestic cheap labor force was sufficient, but at present, China's aging degree is equivalent to Japan's 90s, gradually moving into the aging process, and the demographic changes lead to the rise of labor costs, which will force enterprises to continuously increase their investment in automation and intelligence.
The global manufacturing industry is gradually moving to China, driving the comprehensive development of China's machine vision from hardware to software, upstream components in the light source, industrial cameras are the first to gradually localize, the lens, the image of the underlying software has its survival space. Domestic solution providers have more opportunities to serve head customers.Foreign countries still occupy high-end customers, and three scenarios may occur for Chinese companies.
In the high-end field, the leading domestic enterprises will go head-to-head with Cognex, Keenes, etc., and gradually solidify their industry share through high cost performance after technical catch-up.
In the demand for small and medium-sized enterprises or non-standard products, domestic enterprises have the advantages of high cost performance and local services, but there is the risk of not being able to expand on a large scale because of the difficulty of replicating products and models.
With the development of new technologies such as deep learning and 3D vision technology, there are opportunities for Chinese companies to bend and overtake in new technologies.
Sensing Part: Some applications are shifting from traditional PC-based systems to highly integrated smart cameras
Smart cameras are a collection of traditional PC-based systems with high integration, low cost, and flexible application characteristics, and are currently being developed or launched by machine vision solution providers, industrial camera manufacturers, and automation equipment manufacturers.
Sensing Part: Moving from 2D vision to 3D vision
3D vision can help broaden the fields that 2D vision cannot reach and raise the overall ceiling of the industry: compared with 2D vision, 3D vision can better fuse multiple sensors, detect fast moving targets and obtain depth information such as shape, contrast and spatial coordinate information, and the application areas are much more than 2D vision; international giants and domestic listed companies have taken 3D vision as one of the important R&D directions in the future. The primary market is becoming more and more active in related investment and financing. According to MarketsandMarkets, the global 3D vision sensor market will reach $10 billion by 2025, achieving a compound annual growth rate of 27.6%.
Algorithm Part: From the traditional machine learning into deep learning stage
Deep learning combines the efficiency and robustness of machine vision with the flexibility of human vision and can accomplish detection in more complex environments. For randomly appearing complex appearance detection, traditional machine learning cannot guarantee its stability, while deep learning is a subversion of traditional machine learning algorithms. Deep learning will be integrated into smart cameras and 3D vision. Cognex has launched In-Sight D900, an industrial smart camera with integrated deep learning, in April 2020.
Network Part: The interconnection of everything, the future of industrial interconnection
Machine vision, as the brain and eyes of intelligent manufacturing, is an important data portal for the edge layer of the industrial Internet. The trend of industrial interconnection requires the production control system to be integrated into the upper layer of manufacturing management system and eventually access to the enterprise management system to realize the penetration of information from the production site to the management level. Enterprises in the industry, industry associations and industrial alliances are continuously cooperating and investing in machine vision interconnection and developing basic common standards such as data interface and communication protocols.
CIOE 2021 - Precision Optics, Lens & Camera Module Expo will be held on September 1-3, 2021, covering the latest optical components, lens, camera module, and industrial applications. Concurrent CIOEC Optics Forum will focus on the latest development and trends in optics industry. TSR has already confirmed its presentation in our optics forum. Meet them and other industrial professionals onsite at Shenzhen World Exhibition and Convention Center in September.
Follow us on social media to get insights of China's optoelectronic market.
LinkedIn: @CIOE (China International Optoelectronic Exposition)
YouTube: China International Optoelectronic Exposition (CIOE)
Scan to follow CIOE English WeChat