CN112399115A

CN112399115A - Working mode for improving dynamic range of small-size pixel image sensor

Info

Publication number: CN112399115A
Application number: CN201910752457.0A
Authority: CN
Inventors: 徐江涛; 徐亮; 高静; 赵彤; 查万斌
Original assignee: Tianjin University Marine Technology Research Institute
Current assignee: Tianjin University Marine Technology Research Institute
Priority date: 2019-08-15
Filing date: 2019-08-15
Publication date: 2021-02-23

Abstract

A working mode for improving the dynamic range of a small-size pixel image sensor adopts a time oversampling technology, namely, each pixel in the image sensor is sampled and read for many times; the pixel adopts conditional reset, namely when the pixel is exposed, the pixel is sampled according to a certain sampling interval, when the pixel reaches a set threshold value, the pixel is reset when next sampling arrives, and if the pixel does not reach the threshold value, the pixel continues to accumulate optical signals; each sampling time of the pixels is different, and the threshold value of comparison of each sampling time of the pixels is adjustable; the invention adopts a working mode of combining the conditional reset and the time oversampling technology, the sampling threshold value and the sampling time of each pixel are adjustable, the maximum unsaturated light intensity in the dynamic range of the image sensor is not limited by the capacity of a full well any more, and only depends on the oversampling strategy, thereby realizing the image with high dynamic range.

Description

Working mode for improving dynamic range of small-size pixel image sensor

Technical Field

The invention belongs to the field of CMOS image sensors, relates to an image sensor with small-size pixels and limited full-well capacity, and particularly relates to a working mode for improving the dynamic range of the small-size pixels.

Background

In recent years, according to moore's law, CMOS process technology is continuously improved, so that the pixel size in the field of image sensors is continuously reduced, and the advantages of CMOS image sensors in terms of integration level, power consumption and the like are more and more obvious compared with CCD image sensors.

The dynamic range of the CMOS image sensor is determined by the ratio of the maximum unsaturated signal to the minimum detectable light intensity, and each pixel in the common CMOS image sensor is only subjected to exposure sampling once, so that the maximum unsaturated light intensity is limited by the full well capacity of the pixel; the minimum detectable intensity is related to the noise of the image sensor. With the development of small-sized pixels, the full-well capacity is continuously reduced, so that the dynamic range and the light sensitivity of the CMOS image sensor are low.

Aiming at small-size pixels, in order to overcome the limitation of over-small capacity of a full well and improve the dynamic range of an image sensor, the currently proposed main solution is a multi-exposure technology for carrying out different lengths of exposure time on a target scene. This is therefore at the cost of requiring a complex readout signal chain including column-level amplifiers, associated multiple samples, column-level ADCs, etc., resulting in excessive power consumption.

Disclosure of Invention

Aiming at the problems in the prior art, the invention provides a working mode for improving the dynamic range of a small-size pixel image sensor, which is based on a working mode of conditional reset and time oversampling, aims to break through the limit of full-well capacity of the small-size pixel, overcomes the defects in the prior art, and aims to improve the dynamic range of the image sensor.

A working mode for improving the dynamic range of a small-size pixel image sensor comprises the following main technical scheme:

1) the image sensor adopts a time oversampling technology, namely, each pixel in the image sensor is sampled and read for a plurality of times, the pixel sampled every time is called as a sub-pixel, and the final image is reconstructed according to the sampling result every time, so that the full-well capacity of each equivalent pixel in the final image depends on the time oversampling times m and the threshold value Uth set in each sampling_iThe final equivalent full well capacity is:

；

2) the pixel adopts conditional reset, namely when the pixel is exposed, the pixel is sampled according to a certain sampling interval, when the pixel reaches a set threshold value, the pixel is reset when next sampling arrives, and if the pixel does not reach the threshold value, the pixel continues to accumulate optical signals; in the method, the time of sampling intervals is fixed, and each sampling time is determined by the number of the sampling intervals due to different setting of threshold values; for the extremely low light intensity, before the pixel does not reach the threshold value, the pixel is not reset after each sampling interval, and the pixel is not reset until the collected pixel signal reaches the threshold value, so that the number of the sampling intervals is larger; for strong illumination, the number of sampling intervals is small under the condition of not changing the size of the sampling intervals;

3) the time of each sampling of a pixel is different and the threshold value to which the pixel is compared at each sampling is adjustable.

Fig. 1 shows the overall architecture of an image sensor with small-sized pixels. The image sensor works as follows: firstly, a pixel array is composed of small-size pixels, and the full-well capacity is small and is used for photoelectric conversion; the converted signals pass through a row gating circuit and a column gating circuit to gate specific pixels and read out to a column level; the subsequent column-level judging circuit is used for judging whether the photoelectric signal in the pixel reaches a threshold value at each sampling interval and generating a reset signal of the pixel, wherein the sampling interval is adjustable, and the threshold value of the pixel is set by the threshold value adjusting circuit; then, a column-level digital buffer is used for converting the parallel read binary digital output into serial data; the high-speed interface circuit is used for outputting the serial digital code to the outside of the image sensor; and the logic control is used for controlling the readout of the pixel digital signals.

The invention adopts a working mode of combining the conditional reset and the time oversampling technology, the sampling threshold value and the sampling time of each pixel are adjustable, the maximum unsaturated light intensity in the dynamic range of the image sensor is not limited by the capacity of a full well any more, and only depends on the oversampling strategy, thereby realizing the image with high dynamic range.

Drawings

FIG. 1 is an overall architecture diagram of a small-scale pixel image sensor;

fig. 2 is a graph of the number of photo-generated electrons in a pixel over time.

Detailed Description

The invention provides a working mode for improving the dynamic range of a small-size pixel image sensor, which comprises the following steps: the limitation of full-well capacity of small-size pixels is broken through by means of conditional reset of the pixels, different sampling time of the sub-pixels and a time oversampling technology, and the dynamic range of the image sensor is improved. The invention will now be described with reference to specific embodiments:

fig. 2 is a schematic diagram illustrating the change of the number of photo-generated electrons in the next pixel with time by conditional reset and time oversampling for the present invention, where the sampling interval is fixed to t, the sampling time is determined by the number of sampling intervals, the dotted curve in the diagram is the change of the total photo-generated electrons in the pixel with time, the solid curve is the change of the actual number of photo-generated electrons due to the conditional reset with time, and the dot-dash line represents the threshold value of the pixel. Assuming that the full well capacity of the pixel is 18 electrons, if the quantization precision of the ADC is 2 electrons for the conventional image sensor, the required number of ADC bits is 3 bits. The threshold Uth of the pixel is set to be 18 electrons constantly, the time oversampling is 2bit, namely 4 times, and the equivalent maximum full trap capacity is 72 electrons at this time. Since the times at which the condition reset occurs are 3t, 5t, and 9t, respectively, the lengths of each sampling time are 3t, 2t, 4t, and t, respectively, and thus the digital output of the image sensor is 3, which indicates that the number of generated photoelectrons is 54. Therefore, the image sensor can effectively improve the equivalent full-well capacity of the pixel under the condition of not changing the size of the pixel, thereby improving the dynamic range of the sensor.

Claims

1. A working mode for improving the dynamic range of a small-size pixel image sensor is characterized in that: the main technical scheme comprises:

；