“As one of the three major innovations of smart phones, 3D ToF depth sensing technology relies on the advantages of small size, low error, direct output of depth data and strong anti-interference, and has become one of the key innovation highlights on smart phones in recent years, such as vivo The innovative NEX dual-screen version launched by the company last year adopted ADI’s ToF technology to achieve zero-light-sensing face recognition and achieved good market performance. In fact, as a new type of visual sensing technology, the demand for 3D ToF depth vision technology in major application markets around the world is also increasing, and many solution examples have been born.
As one of the three major innovations of smart phones, 3D ToF depth sensing technology relies on the advantages of small size, low error, direct output of depth data and strong anti-interference, and has become one of the key innovation highlights on smart phones in recent years, such as vivo The innovative NEX dual-screen version launched by the company last year adopted ADI’s ToF technology to achieve zero-light-sensing face recognition and achieved good market performance. In fact, as a new type of visual sensing technology, the demand for 3D ToF depth vision technology in major application markets around the world is also increasing, and many solution examples have been born.
ToF technology decoding
ToF is the abbreviation of Time of Flight, also known as 3D imaging of time-of-flight method. This imaging technology emits continuous infrared light pulses of specific wavelengths to the target, receives the light signal from the object to be measured through a specific sensor, calculates the flight time or phase difference of the light back and forth to obtain the 3D depth information of the object to be measured, ToF camera The brightness images can be quickly connected through the model.
Compared with the other two solutions of 3D depth vision (structured light and binocular stereo imaging technology), the ToF solution has significant advantages in practical applications. For example, post-processing is not required when calculating the depth of field after the picture is taken, which can avoid delays and save the related costs of using a powerful post-processing system; the ToF ranging scale is flexible, in most cases only the light source intensity, The optical field of view and the pulse frequency of the transmitter can be completed; due to its multiple advantages such as resistance to external light interference, small size, fast response speed, and high recognition accuracy, ToF has become the first choice for 3D vision whether it is mobile or in-vehicle applications. Technical solutions. Currently, ToF technology has a wide range of applications, including:
Consumer electronics: virtual reality, face recognition, somatosensory interaction, etc.
New retail: gesture recognition, passenger flow statistics, behavior recognition, etc.
Intelligent security: face recognition, behavior analysis, etc.
Industrial-grade automation: automatic obstacle avoidance, distance measurement, perception positioning, etc.
Medical electronics: augmented reality, remote interaction, etc.
Automotive electronics: assisted driving, face recognition, gesture recognition, etc.
Analysis of Typical Application Cases of ADI ToF Technical Solution
ADI’s 3D ToF technology solution with sensor and front-end chip ADDI903x as the core has ultra-high frame rate and real-time performance. Compared with the point-by-point scanning of 3D laser sensors, it has lower power consumption and footprint, and can be fast and accurate Obtain the depth information of the overall image to achieve VGA image resolution and accuracy. There are a large number of successful applications around the world, and it is being adopted in more and more products. Envision Intelligent ToF modules and solutions are among them Typical representative.
Complete system composition of ADI ToF solution
The ADDI903x series of CCD ToF front-end chips is one of the core components of the ADI ToF solution. This series of products can support CCD infrared light ToF sensors with a resolution of up to 640×480. ADDI903x can convert the image signal into a digital signal, and provide a high-precision pulse time controller. The closed-loop design makes the pulse width controlled by the laser diode more accurate, so that more accurate depth data can be obtained.
“As a company with TOF stereo vision, AI, and edge computing as its core technologies, Envision Intelligence is good at detecting, classifying, identifying and tracking multiple targets such as faces, vehicles, pedestrians, and behaviors. It is also a member of the ADI strategic alliance. Using the ADI ToF solution, combining artificial intelligence, edge computing technology, etc., Envision Technology provides examples of terminal products and application solutions for major mainstream industries.” Envision Intelligence CEO Li Tan said.
01 Passenger flow statistics
The demand for passenger flow statistics is widely present in various scenarios, such as buses, rail transit, chain stores, buildings, scenic spots, etc. How to effectively use imaging technology to accurately distinguish the height and weight of people entering and leaving, while the error of entry and exit time and height is less than 1% , This is quite a technical threshold. The ADI ToF solution is different from the traditional 3D people flow solution. Most of them need to install at least two stereo cameras, only one ToF camera lens mount, which is installed above the door frame and has no installation height limit.
Envision Intelligence innovatively applied the stereo vision module based on the ADI ToF solution to the scene of rail transit turnstiles. The opening and closing control of turnstiles has been realized in the first stage. In addition to user passenger flow statistics, it can also intelligently identify the phenomenon of following fare evasion. When multiple people enter the gate closely, the gate will still be closed, and there will be related warning lights around. The second stage will realize the function of authenticating the user by swiping the face.
02 Gesture/Bone Extraction Algorithm
In terms of human-computer interaction, ToF technology also has a broad application space, and feature recognition technologies such as gestures and bones are widely used in consumer electronics, automobiles and other fields. Compared with traditional solutions, the vision intelligent modular solution based on ADI ToF technology has the advantages of good adaptability, high accuracy and easy deployment. The combination of ToF technology and deep learning algorithms that can run on embedded terminals in real time can obtain in-depth information. On this basis, the background and the target to be measured can be separated, reducing the interference caused by complex scenes and improving the accuracy of recognition.
Conventional detection methods include ultrasonic sensors, infrared sensors, laser sensors, vision sensors, etc. Because ToF can capture images and measure distances, it also has great advantages in scenes such as robot avoidance and navigation. Based on the ToF module developed by ADI, the Vision Intelligence combines the image sensor and VGA ToF sensor module with a built-in image processor solution, which has a better detection angle than traditional sound wave detection, and can accurately measure the distance between the object and the car. Therefore, it can provide a wider range of collision detection prevention for applications such as car reversing systems, door opening protection systems, parking assist systems, and blind spot detection.
When a pedestrian or an obstacle is detected approaching, the ToF module can Display the size and relative distance of the obstacle with an image, as well as a sound warning.
For solution providers in the industry, whether it is for consumer or professional applications, 3D ToF technology still needs to solve a lot of basic problems at present. How to use technical means to truly achieve cost, power consumption, volume, speed, The balance of life, stability, and anti-interference ability has reached a level that is more optimized than the current one, and then the reliability of the actual application of ToF visual sensing technology has been doubled. It is a company such as ADI. The key points that technical solution providers should consider are also a prerequisite for the popularization of ToF technology and the healthy development of the entire market. From the ToF application case of Envision Intelligence, it is not difficult to see that ADI’s ToF technology is implementing faster and more accurate ranging functions in IoT modular system-level software and hardware solutions to help customers achieve simple development and reduce development costs. .