The application technologies of artificial intelligence mainly include voice technologies, visual technologies, natural language processing technologies, and basic hardware, etc. Among them, machine vision technology is a prerequisite for the autonomous operation of robots. It can enable computer systems to observe, recognize, and judge the external environment, and it is one of the most important frontier branches within the scope of artificial intelligence.

Machine vision technology involves many fields such as machinery, electronics, optics, automatic control, computer science, image processing, and pattern recognition. It accounts for more than 40% of the application technologies of artificial intelligence enterprises at home and abroad and has a very wide range of applications.

With High Growth Potential

The global market size of machine vision increased from 3.17 billion US dollars in 2010 to 10.7 billion US dollars in 2020, with a compound annual growth rate of 14.47%. In terms of regional distribution, the European region has the largest market size of machine vision, accounting for 36.4% of the global market; North America and the Asia-Pacific region account for 29.3% and 25.3% respectively, and South America, the Middle East, and Africa account for 9.1%.

Benefiting from the transfer of the global manufacturing center to China, the Chinese machine vision market has developed rapidly and has become another important target market for international machine vision manufacturers outside of Europe, North America, and Japan.

From 1990 to 2000, China's machine vision industry was in its embryonic stage. In the 1990s, machine vision agent enterprises began to appear in China, mainly engaged in the agency services of machine vision devices and technologies. From 2001 to 2010, China's machine vision industry was in its initial stage. Driven by the representative downstream application demands such as the quality inspection of RMB banknotes, the removal of foreign objects in tobacco and raw cotton, and postal sorting, a number of professional machines began to emerge in China. More and more enterprises began to explore and research and develop machine vision software and hardware devices with independent intellectual property rights. As cameras with USB 2.0 interfaces and image capture cards gradually occupied the entry-level market, the product performance continued to improve.

Since 2013, the state has successively issued relevant policies and plans to support the machine vision industry. The "13th Five-Year Plan for Intelligent Manufacturing" issued in 2016 proposed ten key tasks to accelerate the development of intelligent manufacturing equipment and promote the intelligent transformation of key fields. The development plan provides strong support for the development of high-end equipment manufacturing and automated production, and the demand for machine vision products as supporting equipment has gradually increased.

The "Guiding Opinions on Accelerating the Cultivation of New Modes and New Business Forms of Shared Manufacturing and Promoting the High-Quality Development of the Manufacturing Industry" issued in 2019 proposed to support platform enterprises to actively apply artificial intelligence technologies, develop intelligent inspection functions, and continuously improve the intelligent level of the entire manufacturing process.

Machine vision technology provides convenience for intelligent inspection, improves efficiency, and reduces labor costs. With the development of the industry, as customers become aware of the wide applications, efficiency enhancement, and cost reduction advantages of machine vision technology, the penetration rate of machine vision in various industries has gradually increased.

From 2015 to 2018, the market size of China's machine vision industry grew rapidly, remaining above 40%. According to the forecast of the Zhongshang Industrial Research Institute, the market size of China's machine vision in 2020 was 11.5 billion yuan, with a year-on-year growth rate of 11.65%. The compound annual growth rate from 2015 to 2020 was 37.97%, which was 24.31 percentage points higher than the compound annual growth rate of the global machine vision industry from 2015 to 2020.

According to data from CB Insights, in 2020, China has become the world's third-largest market after the United States and Japan.

How to "make profits" in the industrial chain?

The success of a machine vision system depends on the key components used.

In the industrial chain of the machine vision industry, the upstream hardware part includes light sources, lenses, industrial cameras, image processors, and image capture cards; the software includes image processing software and underlying algorithms.

The middle reaches are equipment manufacturers and system integrators.

The downstream applications are extensive, including industries such as semiconductors, automobiles, packaging, pharmaceuticals, and industrial robots.

In machine vision products, the upstream software and hardware together account for 80% of the total cost, accounting for 45% and 35% respectively, and assembly integration and maintenance account for 15% and 5% of the total cost respectively. Therefore, the upstream of the industrial chain is the link with the highest added value in the entire industry.

Light source illumination is one of the important factors affecting the input of a machine vision system, which directly affects the quality of the input data and the application effect. Machine vision light sources mainly use LED lighting, mainly because of its high shape freedom, long service life, fast response speed, good monochromaticity, various colors, high comprehensive cost performance, and other characteristics, which are widely used in the industry.

In 2020, the market size of China's LED lighting was 862.7 billion yuan, with a year-on-year growth rate of 14.30%, and the compound annual growth rate from 2017 to 2020 was 10.71%. In terms of the LED lighting industry, enterprises with an annual sales revenue of more than 2 billion yuan are in the first echelon, including Philips, Opple Lighting, NVC Lighting, Foshan Lighting, Sunshine Lighting, and other domestic and foreign enterprises; the second echelon has an annual sales revenue of 500 million to 2 billion yuan, including enterprises such as Cnlight, Kingsun Optoelectronic, etc.

The main function of the lens is to image the target on the photosensitive surface of the image sensor, which is an important imaging component for image acquisition in machine vision.

After years of layout, business accumulation, and technological upgrading in the lens field, leading enterprises such as Leica, Schneider, Nikon, and Fujifilm have formed leading advantages globally. Domestic enterprises started late, and before 2008, the domestic optical lens market was monopolized by German and Japanese brands.

With the rise of the industry, domestic enterprises in China's industrial lens industry have developed rapidly, mainly focusing on the low-end market, competing with overseas brands with high cost performance. In the high-end market, China still relies on imported high-end products, mainly products from established manufacturers in Japan, Germany, and other countries.

Industrial cameras are a key component of the machine vision system, and the image sensor is the core of the camera. According to the chip type, it can be divided into CCD (Charge-Coupled Device) and CMOS (Complementary Metal-Oxide-Semiconductor) image sensors.

Currently, the global industrial camera industry is dominated by European and American brands. In 2018, North American brands accounted for 62% of the global industrial camera market, and European brands accounted for 15%.

China has an obvious advantage in the CMOS field. By 2014, GalaxyCore had grown into the largest CMOS image sensor supplier in China and has maintained this honor ever since. In 2014, the shipment volume of GalaxyCore's CMOS image sensors exceeded 940 million pieces, and the shipment volume of display driver chips exceeded 100 million pieces, with the annual total sales exceeding 350 million US dollars.

According to the shipment volume statistics, in 2020, GalaxyCore shipped 2.04 billion CMOS image sensors, accounting for 29.7% of the global market share, ranking first in the industry and leaving behind established well-known manufacturers such as Samsung and Sony.

In 2020, the density of industrial robots in China was 246 units per 10,000 people, an increase of 59 units per 10,000 people compared with 2019, rising from the 15th place globally to the 9th place. The density of industrial robots in China is already higher than the global level, but it is still lower than that of developed countries. Among them, South Korea, Singapore, and Japan have long occupied the top positions in the density of industrial robots in countries around the world, with 932 units, 605 units, and 390 units per 10,000 people respectively; the densities of industrial robots in Germany and the United States are 371 units and 255 units per 10,000 people respectively.

Compared with developed countries, there is still a large room for improvement in the density of industrial robots in China, which will drive the demand for automated production equipment, and the demand for machine vision in the segmented fields will also increase. In addition, the cost of upstream components in the industrial chain accounts for 80% of the total cost, mainly focusing on the low-end market, and still fails to break the monopoly of overseas enterprises in the high-end market, which will be an important direction for "making profits" in the future.