From an article in Tech HQ
The Face ID module projects a patterned point cloud of more than 30,000 infrared dots onto surfaces that are roughly an arm’s length (25 – 50 cm) away from the device. If the target surface were perfectly flat then the shape of that point cloud would be unchanged, comparing the optically generated pattern emitted from the iPhone or iPad with the image received by the infrared camera element within the Face ID hardware assembly.
Any variations in the surface reflecting those infrared dots back to the camera will shift the points either closer together or further apart. And the deviations can be used to infer the shape and physical characteristics of the object – in this case, the user’s face – being imaged.
In contrast, returning to the differences between Face ID and LiDAR, the sensor found on the rear of Apple’s products emits fewer infrared marker points, comprising a regular grid of 24 x 24 dots. However, each dot is brighter than the Face ID projected points, which gives the LiDAR on the back of handsets and tablets a much great working range up to 5 m.
And not only can the system determine shape information from distortions in the grid seen when the infrared dots are projected onto the scene in front of the sensor, it can also capture depth information based on time-of-flight. The detection of laser pulses emitted by the LiDAR chip is affected by whether reflecting surfaces are closer or farther away from the device’s infrared source.
Patents such as US-20200256669 (view PDF) show that Apple uses a sparse array of single photon avalanche diodes (SPADs) to perform a kind of optical stocktake on the whereabouts of the infrared range-finding LiDAR emissions. And the iPhone maker’s True Depth Face ID design also features innovations that allows the hardware to perceive distance more accurately, albeit on different length scales.
For the complete article CLICK HERE.
