To obtain the PDF version of this report, please see the end of the article
With the accelerated development of intelligence and automation, the number and specifications of sensor applications will usher in a new wave of growth, with image perception being the top priority. .
Driven by smartphones and hardware, sensor applications have developed rapidly in the past decade; entering the 5GAIoT era, the Internet of Everything is gradually realized, and autonomous driving and robotics technologies are close to the singularity of large-scale commercialization, playing the role of sensory and nerve endings. sensors are expected to usher in a new wave of growth.
Since most AI relies on image recognition for decision-making, optical sensors are not only the core entrance for information collection, but in the near future, the metaverse concept will depict a digital virtual world and will also be the converter and converter between the real world and the mirror universe. The main bridges and industry tracks have long slopes and heavy snow.
CMOS accounts for half of the optoelectronic components, driving the overall market to grow at a compound rate of close to 10% in the next five years.
According to IC insights, the OSD semiconductor market size in 2020 is US$88.3 billion (accounting for 20% of the overall semiconductor market), of which optoelectronic components, sensors, and discrete devices account for 50/19/31. The OSD market is expected to reach 20% in 25 years. 132.6 billion, with a CAGR of 8.5.
The market size of CMOS/LED lighting/laser transmitter/coupler/CCD in optoelectronic components in 20 years is 19 billion/13.3 billion/2.2 billion/1.9 billion/1.2 billion respectively. The overall market is expected to be in 25 years It will reach 68.3 billion, with a CAGR of 9.2, mainly driven by the demand for image sensors, machine vision, face recognition, 3D depth perception, autonomous driving images, solid-state lighting, and optical communications.
CMOS is the core component of optical sensing, accounting for nearly half of the cost of mobile phone/car camera modules.
Generally speaking, CMOS/lens/VCM/assembly account for about 40/20/10/15 of the total camera cost respectively. As the core component of optical sensing, CMOS accounts for nearly 50% of the cost of optical systems such as camera modules and lidar.
In the future, driven by high-definition images, high pixels may further increase the cost ratio of CMOS.
The downstream applications of CMOS are gradually shifting from mobile phones to automobiles, industrial security and VRAR.
According to Omida, the proportion of CMOS downstream applications in mobile phones, industrial security and automotive applications in 2020 is 73/11/6 respectively (based on market size); lower than 5MP/6-12MP/13-32MP/ The proportions of 33-50MP/higher than 51MP are 27/32/16/14/11 respectively.
With the rise of autonomous driving (L3 is expected to reach 12 in 25 years), it is expected to replicate the multi-camera upgrade trend brought by smartphones (18 bicycle cameras, 01 lidar).
In addition, with the sales of the cost-effective Oculus 2 and Apple’s launch of MR headsets next year, ARVR is also expected to bring potential growth in the next five years (estimated shipments in 25 years will reach 30 million, 8 units per unit) .
Replicating the growth trend of smartphones, the scale of the automotive market will be comparable to that of mobile phones in the long term.
Currently, bicycle cameras are generally equipped with 1-2 cameras (1 front view and 1 rear view). However, judging from the trend of autonomous driving, a complete ADAS system should generally include at least 6 cameras (1 front view, 1 rear view, 4 surround view), and the number of cameras in high-end smart cars can reach 8-12 indivual.
According to IDC forecasts, global self-driving car shipments will exceed 55 million units in 2024, vehicle camera shipments will exceed 440 million units, and the CAGR from 2019 to 2024 will exceed 20.
At the same time, the value of vehicle-mounted cameras far exceeds that of mobile phone cameras. Although vehicle-mounted cameras have low pixel specifications, their CMOS prices are much higher than mobile phone CMOS due to their safety, and the long-term market size is expected to be comparable to that of mobile phones.
Multi-sensors for autonomous driving are the trend, and lidar is an important addition to optics. Shipments may reach 300 million units in 25 years, and the market size exceeds US$1.5 billion.
Currently, mainstream autonomous driving sensors mainly include vehicle cameras, ultrasonic radar, millimeter wave radar and lidar. As the level of autonomous driving increases, the number of on-board cameras and radars continues to increase.
According to TSR forecasts, the global market size of automotive lidar in 2025 will be approximately US$1.5 billion, with a CAGR of over 80. The current price of lidar is more than US$3,000. As the industry chain matures, the cost is expected to drop to US$500 in 2030. After the price reduction, the advantages of high accuracy and wide detection range will be highlighted, and the market space is expected to further expand.
Tightening international politics and the development of smart cities have driven the rapid growth in demand for high-definition security cameras.
Security cameras are widely used in public services, corporate services and civilian scenarios. According to Yole data, the global security CCM market was US$860 million in 2018, and the market size is expected to reach US$2 billion in 2024, with a CAGR of approximately 15 from 2018 to 2024.
With the development of smart cities and the advancement of new infrastructure, as well as the increase in non-civilian demand brought about by the political situation, the demand for security video surveillance cameras continues to increase.
According to TSR data, global security lens shipments in 2019 were 313 million units, and the CAGR from 2014 to 2019 exceeded 20. At the same time, in response to the call of the Ministry of Industry and Information Technology to “put 4K first and take 8K into account”, high-definition security cameras will be the future trend.
It is expected that the shipment proportion of high-definition cameras (1080P) will reach 76% in 2021, and the proportion of zoom lenses will also gradually increase. A single VR machine requires multiple cameras, and ARVR is expected to continue to grow in the next five years.
Oculus quest2 has 4 cameras distributed around it for position tracking, perspective and hand tracking.
According to the U.S. Patent Office, Apple's patented display VR headset may be equipped with 8 surround-sensing cameras with overlapping fields of view to simultaneously provide a 360-degree physical view of the user's environment.
According to IDC forecasts, global VR headset sales will reach 28.6 million units in 25 years, with a compound growth rate of approximately 40% from 2021 to 2025. If a single machine has 8 cameras, the shipment volume will exceed 200 million.
In terms of AR optics, AR equipment includes multiple cameras, including depth-of-field, environment, and high-definition cameras, many of which are 3D cameras used to obtain information for modeling.
CMOS stands on three pillars, Weir and Gekko are steadily making breakthroughs, while ON Semiconductor and SmartSite are deeply exploring market segments.
The CMOS industry is currently dominated by Sony (47), Samsung (20), and Weir (12). From the perspective of mobile phone CMOS, Sony (48), Samsung (26), and OmniVision (14) still rank among the top three, followed by Gekko Microelectronics (5) and STMicroelectronics (4). Among them, Gekko Microelectronics and STMicroelectronics (4) STMicroelectronics accounts for 60% of the market share below 5MP, while Sony accounts for 80% of the market share from 6MP to 12MP. From 13MP to 32MP, Samsung, Vail, and Sony dominate the market. Above 32MP, Samsung dominates, followed closely by Sony. thereafter. ON Semiconductor and SmartSite have advantages in the automotive and security segments.
The domestic industrial chain may rely on high growth in domestic demand to achieve substitution.
At present, the industry mostly adopts the IDM model, and a small part outsources OEM packaging.
Under the trend of localization, pure design Weir (Howell) and Gekewei have made breakthroughs in the high-end and mid-low-end markets respectively, especially in the automotive and security fields, the domestic supply chain has accelerated its development. We judge that major foundries such as TSMC, Semiconductor Manufacturing International Corporation, Hua Hong Semiconductor, and packaging and testing plants such as Jingcai, Tongxin Electronics, KYEC, Jingfang Technology, etc. are expected to benefit from the increase in outsourcing share and process upgrades.
As a segment that accounts for half of the non-IC market, optoelectronic components are expected to benefit from the continuous penetration of sensing applications.
As the core component, CMOS is expected to be the most beneficial link driven by emerging fields such as autonomous driving and ARVR.
Leading manufacturers such as Vail and Geko Micro are expected to continue to make breakthroughs in technology and share driven by the domestic supply chain; and the unique design foundry model will also support TSMC, SMIC and Hua Hong Semiconductor, as well as back-end packaging and testing materials, Tongxin Electronics, KYEC, and Jingfang Technology have experienced rapid growth in performance.
The penetration of autonomous driving is less than expected, and external factors have caused uncertainty in the manufacturing process
——————————————————
Please pay attention to learn about the latest industry analysis reports every day!
The report belongs to the original author, and we do not make any investment recommendations!
Original name of the report: "Image Sensor: The Bridge between the Real World and the Digital Universe"
Author and Analyst: Hong Jiajun and Li Jiaxun of Industrial Securities
Get more PDF versions For reports, please log in to the official website of Yuanzhan Think Tank or click on the link: "Link"