Pixel structure of a retina-like image sensor

Abstract

A kind of pixel structure of a retinal image sensor, related to the field of integrated circuits. The existing method uses asynchronous real-time detection of light intensity changes for each pixel, and uses the AER event readout method, which is prone to delays, resulting in events and corresponding events. The time stamps do not match, resulting in motion artifacts; and it is susceptible to false events caused by the leakage current of the transistor, and cannot be removed. The adaptive light sensing unit is used to complete the conversion of optical signal to electrical signal and output voltage signal; the sampling comparison unit is used to complete the sampling and comparison function to determine the polarity of the signal. When the light intensity becomes weaker, an OFF signal is generated, and the light intensity becomes stronger. The signal processing unit stores the polarity signal output by the sampling comparison unit, and performs signal processing according to the instruction to complete the noise reduction; the present invention eliminates motion artifacts and abandons the asynchronous readout method of the traditional dynamic visual sensor, Grayscale images can also be output while outputting events.

Description

technical field [0001] The invention relates to the field of integrated circuits, in particular to a pixel structure of a retinal image sensor. Background technique [0002] Traditional image sensors perform imaging in frames. Compared with the previous frame, regardless of whether the pixels have changed, a series of processing such as exposure, sampling, and amplification must be performed on all pixels. There is a lot of repetitive work. High data redundancy. Moreover, due to the limitation of the frame rate, the image information of the high-speed moving object is easily lost in the frame-to-frame interval. [0003] Nowadays, the mainstream imaging devices are CCD and CMOS image sensors. Their pixel structure is simple, small in size, which is conducive to improving image resolution, and output images in frame imaging mode, so that the output images are more intuitive and can better please people. Eye. But the frame-based imaging mode also has the following inherent d...

AI Technical Summary

Problems solved by technology

[0016] In order to solve the existing method, the present invention uses asynchronous real-time detection of light intensity changes for each pixel, and adopts the AER event readout method, which is prone to delays, resulting in mismatches between events and corresponding time stamps, and motion artifacts; And it is easy to be affected by the leakage current of the transistor to generate false events, and it cannot be removed, etc., providing a pixel structure similar to the retinal image sensor

Images