In low light (usually the illuminance is 10-1lx lower than the illuminance of the full moon night sky), the amount of light that the human eye can receive is very small. The quantum fluctuations of this time have an important impact on human vision.

·

Radiation is a discrete discrete process, and the light energy of radiation is carried by photons. The photon rate of radiation varies instantaneously, and the number of photons radiated at this moment and that moment is different. That is, photon emission fluctuates over time. Generally speaking, the number of photons received in a unit time is just the average value of the number of photons fluctuating with time. The deviation of the instantaneous value n from the statistical mean N is called "fluctuation" and is expressed by the root mean square deviation. It can be known from probability theory that the probability of photon emission according to Poisson distribution, its root mean square fluctuation value is equal to.

 

The vision of the human eye in low light is the result of the absorption of photons by the retina within a certain time interval (accumulation time) of the eye. Obviously, this is a process that depends on the fluctuation of radiation. The average number of photons received by the human eye during the accumulation time has a "fluctuation" on the average. This "fluctuation" value reduces the ability of the eye to detect the difference in the number of photons between adjacent pixels on the target. In other words, human vision is limited by photon fluctuations in low light. From this point of view, a statistical theory of visual room detection was formed, which is used to determine the detection of objects and serve as the basis for the display image detection capabilities of electro-optical imaging systems.

 

The theory of visual detection in low light was initially developed by Vries and Rose in the early 1940s. Its concept and model are relatively simple. Assuming that the eye absorbs N photons from the scene on average during the accumulation time, the "fluctuation" around this average value is √N. At this time, the ability of the eye to detect the minimum change AN of the N value is limited, then there is

ΔN∞

or

ΔN=K

In the formula, ΔN is the change in the number of photons that can be detected by the eye, which is the detected signal; the "fluctuation" value of photons interferes with the visual detection of the human eye, which is called photon noise; the proportional coefficient K is called the signal-to-noise ratio, K=ΔN/. Only when the K value is greater than the threshold SNR, the signal can be detected, and the threshold SNR is determined experimentally.

 

The development of low-light technology for many years has become very mature. The Sytong HD-syll1 high-definition near-low-light night vision technology independently developed by Shenzhen Shiyutong Technology Co., Ltd. adopts super star-level night vision full-color image NIR, and the advanced technology is greatly improved Quantum efficiency, sensitivity increased by 31%, read noise decreased by 16%, and dark current decreased by 46%. Sytong HD-syll1's original back-illuminated structure, by illuminating the back of the silicon circuit board, can avoid the influence of wiring and transistors, increase the amount of light entering a unit pixel, and at the same time, it can also suppress the decrease in sensitivity caused by the change of the light incident angle. The problem. The sensitivity is twice that of the traditional front-illuminated CMOS image sensor, and it achieves low noise. The advanced shutter effectively increases the sensitivity and reduces noise. It can also shoot high-definition, smooth, and high-quality images in extremely dark scenes at night. .