Application Breakout Optical Hardware Boom Warms Up

3D visual perception technology was first used in the industrial field, mainly for high-precision three-dimensional measurement and small deformation measurement of objects and materials, etc. 3D visual perception is mainly used in three-dimensional scanning, small deformation measurement, bending angle measurement and analysis, positioning and navigation of industrial robots in the industrial field. At the end of 2023, Obi Zhongguang used large models of speech, language, vision-language, supplemented by data input from the company's Orbbec Gemini 2 series of depth cameras, to create a robotic arm capable of understanding and performing voice tasks.
The lead of Apple mobile phone makes 3D vision sensor in the field of mobile phone scale application, but also marks the 3D visual perception technology in the field of consumer scale began to popularize.
In addition to brush face payment, 3D vision sensors are also accelerating landing in areas such as smart door locks and 3D house viewing. Based on 3D visual perception in consumer electronics, finance, retail, catering, automotive, AIoT and other industries landing applications, such as biometrics, 3D reconstruction, skeleton tracking, AR interaction, digital twins, autonomous positioning and navigation. The global 3D visual perception market will grow rapidly and is expected to reach $17.2 billion billion by 2028.
Compared with traditional machine vision, robot vision needs to have the characteristics of 3D vision, high integration, and complex and changeable scenes. The strong demand will promote the rapid evolution and iteration of various mainstream 3D visual perception technologies and promote the accelerated development of the robotics industry.
Sensors in the field of automatic driving are divided into visual routes and lidar routes. Visual solutions are derived from perceptual components such as cameras and radars. Cameras are the main components, supplemented by sensors such as millimeter wave radar and ultrasonic radar. The overall cost is relatively low, but the visual scheme has extremely high requirements on algorithms. The lidar scheme uses lidar and cooperates with sensors such as cameras, millimeter wave radar and ultrasonic radar, emphasize "strong perception, low algorithm".
With the continuous improvement of L2, L3 and above penetration, the number of sensors carried by single vehicle will increase significantly compared with the past, whether it is visual route or lidar route. The number of L3 single vehicle sensors is expected to reach 17-34, which is a significant increase compared with L1's 6-14. According to Dongguan Securities, by 2025, the market size of domestic passenger car automatic driving sensors is expected to be about 52.77 billion yuan, of which the market size of on-board cameras, millimeter wave radar and ultrasonic radar is expected to reach 13.294 billion yuan, 14.22 billion yuan and 17.241 billion yuan respectively.