At present, 3D face recognition in the market is mainly divided into: 3D structured light, ToF camera, binocular stereo vision according to the imaging principle of the camera.
3D structured light
3D structured light projects light with certain structural characteristics onto the object to be photographed through an infrared light projector, and then is collected by a special infrared camera. The principle of similarity of triangles is mainly used for calculation, so as to obtain the depth information of each point on the image, and finally obtain three-dimensional data. Face recognition based on 3D structured light has been practically applied on some smartphones.
ToF camera
The illumination light source of ToF camera generally adopts square wave pulse modulation, and the distance is measured according to the time difference between pulse emission and reception. The 3D data obtained by the ToF method can be used for smart door locks, people flow statistics, VR&AR, etc. The purpose is to use it as an interactive device with the machine, used in games. The main acquisition and processing is the attitude data.
Binocular stereo vision
Binocular is a method based on the principle of parallax to obtain three-dimensional geometric information of objects from multiple images. Two digital images of the object to be measured are simultaneously obtained by dual cameras from different angles, and the three-dimensional geometric information of the object is recovered based on the principle of parallax, so as to obtain the depth information of each point on the image, and finally obtain three-dimensional data. Because the principle of binocular stereo vision imaging has relatively high requirements on hardware, especially the focal length of the camera and the plane position of the two cameras, the application range is less than that of 3D structured light ToF.
