Foldable Electric Scooter,Folding Scooter,Folding Electric Scooter,Foldable Scooter Ningbo Hiley Technology Co.,Ltd , https://www.hiley-mobility.com
Depth camera application and its main technology
Apple has always been technology-driven innovation, and every year it introduces technological innovations at the time of its hardware release. For 17 years, on its iPhone X, the 10th generation iPhone, we saw face recognition unlock: Face ID.
In fact, Apple acquired PrimeSense in 2013, and has been working on miniaturizing the depth camera of PrimeSense. The innovation in the iPhone X is the use of a front depth camera. Now, the application of depth camera has been unlocked by intelligent terminals, and gradually extended to mobile payment, live detection and other scenes. The technical means behind it: 3D structured light has gradually entered the public eye.
Structured Light : A collection of projected rays of known spatial orientation is called structured light. 3D, as the name suggests, projects in three dimensions. Three-dimensional imaging based on structured light is actually the measurement and reproduction of three-dimensional parameters, which is mainly distinguished from pure passive three-dimensional measurement technology such as binocular stereo vision. It is necessary to project structured light onto the measured object through structured light. The deformation (Structured Light, or Time of Flight - ToF) determines the size parameter of the object under test. The 3D structured light technology mainly emits light with certain structural features emitted by a near-infrared laser, and is projected onto the object to be photographed, and then collected by a special infrared camera. Since the light with a certain structure is in different depth regions of the object, and the image generated after the acquisition changes with respect to the original light structure, and then the calculation unit converts the change of the structure into depth information, we obtain The three-dimensional structure of the object being photographed.
At present, the 3D structured light scheme is mainly applied to the unlocking of high-end smart models. The 3D structured light imaging is safer than the 2D face unlocking, and the technical threshold is naturally higher. There are not many companies in the market that have the ability to self-develop 3D structured light technology.
The entire system of 3D structured light includes structured light projection equipment, cameras, image acquisition and processing systems. The process is that the projection device emits light onto the object to be measured, and the camera captures a three-dimensional light pattern formed on the object to be measured, and the captured image is processed by the acquisition processing system to obtain surface data of the measured object. In this system, when the relative position of the camera and the projection device is constant, the degree of distortion of the light projected on the object to be measured depends on the depth of the surface of the object, so that a light image having a depth can be obtained in the captured image. The 3D structured light mode contains the pattern of points, lines, and faces, and refers to the type of light projected. In general, a network structure consisting of a plurality of vertical bidirectional lines is commonly used, and this mode can realize three-dimensional contour measurement without scanning, and is fast.
The encoding of the projected pattern
The pattern projected by structured light method needs to be carefully designed and coded. There are many ways to structure optical coding, which are generally divided into the following categories:
1, direct coding (direct coding)
Depending on the image grayscale or color information encoding, a wide spectral range is required.
Advantage: All points are coded, theoretically achieving higher resolution.
Disadvantages: It is greatly affected by environmental noise and the measurement accuracy is poor.
Time multiplexing coding
As the name implies, the technical solution needs to project N consecutive sequences of different coded lights, and the receiving end identifies each code point according to each of the N consecutive sequence images received. The projected coded light has binary code (common), N-ary code, grayscale + phase shift and other schemes.
The advantages of this solution: high measurement accuracy (up to micron level); higher resolution depth map (because there are a large number of 3D projection points); very small color effect (using binary encoding).
Disadvantages: It is more suitable for static scenes, not for dynamic scenes; the amount of calculation is large (because it is necessary to calculate continuous N projections for identifying a code point).
Spatial multiplexing coding
The code is identified based on the distribution of all points in a window within the surrounding neighborhood.
The advantage of this technology: suitable for moving objects.
Disadvantages: Discontinuous object surfaces may produce erroneous window decoding (because of occlusion).
Derivative application of 3D structured light
· 3D anthropometry
The human body data is scanned through the 3D structured optical module, and the result can truly reflect the posture and physical information of the subject. It is reported that 3D structured light is used to measure height, arm length and other information in a few seconds, the error can be controlled in millimeters, and can be used for human body modeling, physical identification and other purposes.
· Object volume measurement
3D structured light can measure depth information. For the logistics industry, 3D structured light can be used to quickly identify the long, wide and high information of objects, so logistics companies can better regulate storage management and make rational use of storage and transportation space. Improve space utilization.
· Physical 3D modeling
The depth camera can be used for environment sensing, 3D scanning, SLAM, gesture recognition, etc., and enhances the fusion and interaction of virtual reality in the AR/VR field. The device in the figure can be understood as an additional structured optical module of the touch screen integrated machine, which is convenient for designers. Get three-dimensional images and information of objects.
· Face 3D modeling
3D structured light can be used to model the human body and the face in real time. For the face recognition and control scene, the influence of factors such as facial occlusion can be reduced, so it has broad application prospects.
· Face recognition payment
At present, the fingerprint payment function is very mature, and face recognition is mainly applied to payment identity verification and login identity verification to supplement.
Summary of 3D structured light scheme
1, advantages
· Because structured light actively projects coded light, it is ideal for scenes with insufficient illumination (even no light) and lack of texture.
· Structured light projection patterns are generally designed to achieve high measurement accuracy within a certain range.
· Technology is mature, and depth images can achieve relatively high resolution.
2, disadvantages
· The outdoor environment is basically unusable. This is because it is susceptible to strong natural light outdoors, causing the projected coded light to be submerged. Increasing the power of the projected light source can alleviate the problem to some extent, but the effect is not satisfactory.
· The measurement distance is close. The farther the object is from the camera, the larger the projection pattern on the object and the worse the accuracy (imagine the situation where the flashlight is shining far away), and the corresponding measurement accuracy is also worse. Therefore, the measurement accuracy of the depth camera based on structured light is greatly reduced as the distance increases. Therefore, it is often applied more in close-range scenarios.
· Easy to be affected by smooth surface reflections.
The 3D structured light scheme module tends to be miniaturized, and has high requirements for chip design and production process. Therefore, there are not many companies that have independent 3D structured light technology and put into application: Huawei, Apple, Obi Zhongguang, intel (RealSense), Microsoft (kinect).
At this stage, it is mainly used for unlocking and identity check on high-end smart models. The accuracy and speed of unlocking/recognition are better than fingerprint recognition. In terms of virtual interaction, it has also been used for somatosensory games, gesture control, and the like. The technical conditions for 3D structured light landing in the application scenario are already available, and more derivative application scenarios are being developed and tried.