Depth Imaging Using 830nm/940nm Infrared Speckle Patterned Light

In September last year, Apple Inc. of the United States released the smartphone iPhone X. For the first time, a front-facing depth camera was installed on the smartphone to support face 3D data acquisition and implement face-based unlocking based on 3D information. After dismantling the analysis by industry insiders, it was learned that it used infrared speckle random structure optical imaging technology.

The recent exposure to a company dedicated to deep recognition technology research and development - Guangzhen Zhishen, the core technology of its research is to use 830nm / 940nm infrared random speckle structure light to achieve 3D imaging. The current team has developed two solutions, one is chip-based 3D imaging, and the other is 3D imaging based on pure software algorithms.

Guangzhen Zhishen believes that 3D information is being used in smart phones, industrial measurement, unmanned vehicles, robotics, surveillance industry, and augmented reality. According to Li Zhihua, the company’s marketing director, “there are a few mobile phones that implement the function of “brushing the face” at present, and all adopt the traditional flat camera. In comparison, the 3D camera can better perceive three-dimensional information because it collects the vertical distance information at the same time. Space, the “flat” type of fraud for photographs, videos, etc. can achieve natural anti-counterfeiting, so the security level is higher; at the same time, the information it collects is more abundant, and it performs better than traditional cameras in misrecognition rate and refusal rate. "It is worth mentioning that the 3D camera adopts near-infrared laser active projection to adapt to the dark environment, so using the phone at night can also achieve facial recognition.

940nm LED and 850nm LED are all widely use in ir camera. Best LED factory will be able to package this IR LED with SMD LED type and LED Lamps type. The imaging effects from left to right are as follows: Guangzhen Zhishen, Microsoft Kinect (Structural Light Technology), and domestic founder and CEO of Jing Pinzheng Zhishen, Yu Dayong, told 36, the current scheme of Guangzhen Zhishen pure software algorithm. Has many outstanding advantages, including: - Fast imaging, Android mobile phones reach 30fps; - Optical distortion tolerance of 5%, can reduce the optical device process requirements, improve production yield, and resistance to mechanical deformation that is anti-fall; - Assembly The process does not need to be adjusted to reduce labor costs and equipment costs to ensure mass production capacity; - Better sub-pixel estimation improves Z-axis measurement accuracy; - Fast initial calibration algorithm, one frame of image time can complete all deformation, temperature calibration, and imaging calculations ;- The same algorithm can be used for both active monocular and active binoculars, all capable of high-quality imaging; - In the mobile phone face unlock scene, fast, can provide a better user experience; - When integrated into a smart phone, because No separate chip is required. This solution has the outstanding advantages of low cost and small size.

Guangzhen Zhishen also has chip-based imaging solutions that can provide customized solutions based on market characteristics.

Guangzhen Zhishen is currently in contact with well-known optical module manufacturers in China and has produced a depth camera Demo for smartphones. He has started contact with a number of mobile phone manufacturers and will use the combination of "Zhenzhen Zhishen + optical module manufacturers". In the form of a 3D camera for smartphone manufacturers. According to Canalys data, the total smartphone shipments in China in 2017 were 459 million. With the increasing demand for 3D face recognition from smartphones, the demand for 3D cameras has become increasingly strong. At the same time, Guangzhen Zhishen will also be involved in industrial inspection, civil access control systems and other fields. According to the calculation of 10 million mobile phones per year, the annual mobile phone depth camera module will exceed tens of millions of yuan per year.

Since Zhenzhen Zhishen provides software algorithms for the depth camera of mobile phones, there is basically no need for additional costs. In the future, Guangzhen Zhizhi hopes to extend and expand 3D modeling and printing, industrial assembly line inspection, military/public access control systems, AR applications, somatosensory control, gesture recognition, and intelligent logistics monitoring. Yu Dayong said: “These are all deep recognitions. In areas where technology fits, in these areas we will provide integrated solutions that focus on in-depth identification and integrate upper-layer hardware and software applications.” Guangzhen Zhishen completed the Angels round of financing in 2016. The predecessor of the team was Tsinghua University Tianjin Electronics. The Informatics Research Institute's in-depth identification project team, after years of technical accumulation, has achieved numerous results in depth measurement, human posture recognition, and gesture recognition, and has formed a complete solution. In order to achieve a better technology landing, the team created Zhizhen Zhizhi deep and hopes to apply deep visual perception technology to more fields.