Image sensor dark current inter-frame fluctuation measurement method removing the influence of random noise

Abstract

The application discloses a method for measuring dark current interframe fluctuation of an image sensor by removing random noise influence, which comprises the following steps: obtaining M*N image information of I frames by the image sensor, taking average of the gray value of each frame of image data in the frame, and recording the average as a dark current value of each frame; calculating the standard deviation of I dark current values, and recording the standard deviation as a dark current interframe fluctuation value BLC_stability; averaging the gray value matrix of I frames of image data to obtain an average gray value matrix; subtracting the average gray value matrix from the gray value matrix of any frame of image data to obtain a gray value matrix containing only random noise; calculating the standard deviation of the gray value matrix containing only random noise, and recording the standard deviation as random noise sigma RN ; and calculating a dark current interframe fluctuation value removed from the influence of random noise. The application can solve the problem that the traditional dark current interframe fluctuation value contains residual random noise interframe fluctuation, and realize more accurate measurement of the dark current interframe fluctuation value.

Description

Technical Field

[0001] This invention relates to the field of CMOS image sensors in integrated circuits, and more particularly to the field of methods for testing dark current and noise in image sensors. More specifically, it relates to a method for measuring inter-frame fluctuations in dark current of image sensors by removing the influence of random noise. Background Technology

[0002] CMOS image sensors, due to their advantages such as low power consumption and high integration, have gradually replaced CCD image sensors in the market. After decades of development, CMOS image sensors have been widely used in many cutting-edge technological fields such as medical diagnostics, smartphones, autonomous driving, and security monitoring. Noise, sensitivity, and signal-to-noise ratio are key indicators for evaluating their imaging performance. In recent years, with technological advancements, the exploration of imaging under extreme conditions has become increasingly in-depth, and low-light imaging has become an important direction for image sensor development. The key factor determining the imaging quality under low light is the magnitude of noise, among which dark current noise contributes significantly to the image quality. Due to limitations in manufacturing processes, each pixel cannot be made completely identical, and the magnitude of its dark current cannot be uniform. This not only causes differences in dark current at different spatial locations within the same frame but also leads to fluctuations in the average dark current value between different frames. The fluctuation of dark current between frames reflects the stability of dark current in the time domain. Especially under low-light imaging, its impact on image quality is particularly significant and is one of the important parameters for evaluating the imaging performance of image sensors.

[0003] Traditional methods for measuring inter-frame fluctuations in dark current, such as Figure 2 As shown, 100 frames of images were first acquired in a completely dark environment, and the mean value of each frame was calculated. The fluctuation of these 100 means is as follows: Figure 1 As shown, the standard deviation of these 100 means is then calculated, which is the value of the dark current inter-frame fluctuation. This method assumes that averaging across each frame completely eliminates random noise, and the impact on inter-frame fluctuations is negligible. However, in practical applications, under low-light conditions, the inter-frame fluctuation value of dark current is very small, within 10... -3 The magnitude is on the order of magnitude, and after averaging a finite number of points, the remaining random noise inter-frame fluctuation values ​​are also around 10. -3 The magnitude is so large that it can still interfere with the inter-frame fluctuation values ​​of the aforementioned dark current, thereby affecting the assessment of the true inter-frame fluctuation values ​​of dark current and hindering the investigation of its causes. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and propose a method for measuring the inter-frame fluctuation of dark current in image sensors by removing the influence of random noise. This method can solve the problem that the traditional inter-frame fluctuation value of dark current contains residual random noise, and achieve more accurate measurement of the inter-frame fluctuation value of dark current.

[0005] The objective of this invention is achieved through the following technical solutions.

[0006] The present invention provides a method for measuring inter-frame fluctuations in dark current of image sensors by removing the influence of random noise, comprising the following steps:

[0007] 1) In a dark environment, continuous image capture using an image sensor yields I-frames of image information with a resolution of M×N, where M is the number of pixels in the horizontal direction and N is the number of pixels in the vertical direction. The grayscale values ​​of each frame are averaged over the frame, and this average value is recorded as the dark current value for each frame. The average value of the i-th frame is expressed in μ... t_i This indicates that the value of i ranges from [1, I], and μ t_i Given a one-dimensional array with I elements

[0008]

[0009] In the formula, m and n represent the coordinates of different pixels in the pixel array of the image sensor, with m ranging from [0, M-1] and n ranging from [0, N-1]. t_i This represents the grayscale value of the pixel at coordinates (m,n) in the i-th frame.

[0010] 2) Calculate the standard deviation of I dark current values, denoted as the dark current inter-frame fluctuation value BLC_stability:

[0011] BLC_stability=std(μ t_i )

[0012] 3) Average the grayscale value matrix of the I-frame image data obtained in step one to obtain the average grayscale value matrix y. RN_ave ;

[0013]

[0014] In the formula, m and n represent the coordinates of different pixels in the pixel array of the image sensor, with m ranging from [0, M-1] and n ranging from [0, N-1]. RN_i This represents the grayscale matrix of the i-th frame image data;

[0015] 4) For the I-frame image data obtained in the first step, take the grayscale matrix y of any one frame of image data. RN_i Subtract the average gray value matrix y obtained in the third step from the gray value matrix of this frame image data. RN_ave The resulting matrix is ​​the gray-level matrix y containing only random noise. RN ;

[0016] y RN [m][n]=yRN_i [m][n]-y RN_ave [m][n]

[0017] 5) Calculate the standard deviation of the gray-level matrix containing only random noise obtained in step 4, denoted as random noise σ. RN ;

[0018] σ RN =std(y RN )

[0019] 6) Calculate the dark current inter-frame fluctuation value BLC_stability′ after removing the influence of random noise using the following formula.

[0020]

[0021] Compared with the prior art, the beneficial effects of the technical solution of the present invention are as follows:

[0022] This invention considers the effect of averaging on random noise removal during testing, and the impact of residual noise on dark current inter-frame fluctuation testing. Therefore, in the proposed method for measuring dark current inter-frame fluctuation, residual random noise is removed, improving the accuracy of dark current inter-frame fluctuation testing. This allows for a more precise investigation of the causes of dark current inter-frame fluctuation, thereby reducing the impact of dark current on image quality from the source. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of inter-frame fluctuations in dark current.

[0024] Figure 2 This is a schematic diagram of a traditional method for measuring inter-frame fluctuations in dark current. Detailed Implementation

[0025] The present invention will now be further described with reference to the accompanying drawings.

[0026] To obtain a more accurate measurement of inter-frame dark current fluctuations, the residual random noise can be subtracted beforehand from the traditional measurement. The residual random noise can be calculated as follows:

[0027] In the first experiment, under dark conditions, T frames of image information (x*y) were obtained using an image sensor, where x represents the number of pixels in the horizontal direction and y represents the number of pixels in the vertical direction. After... Figure 2 After measuring random noise using the traditional method shown, the random noise follows a Gaussian distribution with mean μ and standard deviation σ. The average grayscale values ​​of the image data within each of the T frames are calculated to obtain T means, denoted as [the average values ​​are then calculated]. in Let represent the mean grayscale value of the i-th frame in the first experiment, where i ranges from [1, T] and is a set of values ​​that follow the order . A sample with a distributed mean. Following the above process, in the second experiment, another sample with the mean can be obtained, denoted as... in Let represent the mean grayscale value of the i-th frame in the second experiment, where i ranges from [1, T]. This continues until the t-th experiment, yielding... Let Si represent the mean grayscale value of the i-th frame in the t-th experiment, where i ranges from [1, T]. Let S1 be the variance of the E1 sample. 2 The variance of the E2 sample is S2. 2 , ..., E t The sample variance is S t 2 Record S 2 =[S1 2 S2 2 S3 2 , ..., S t 2 Let S[t] represent the set of t variances obtained from t experiments. 2 The distribution it follows is derived as follows:

[0028] Since the sample variance follows a Gaussian distribution, it follows the following distribution:

[0029]

[0030] Chi-square distribution (χ²) 2 From the formulas for the mean and variance of the chi-square distribution, we know that the mean is equal to the degrees of freedom of the chi-square distribution, and the variance is equal to the degrees of freedom * 2.

[0031]

[0032]

[0033] therefore,

[0034]

[0035]

[0036] Therefore, the expected value E of the mean and variance is The variance D is

[0037] Thus, we can obtain the distribution of the residual random noise after each measurement experiment has been averaged, and its expected value is: Considering the influence of residual random noise, the expected amount of residual random noise is used as the value that needs to be corrected in this invention. In order to calculate the expected amount, the process of measuring the random noise value needs to be added to the traditional measurement method.

[0038] Based on the above theoretical proof, this invention proposes a method for measuring inter-frame fluctuations of dark current in image sensors that removes the influence of random noise. Considering the residual random noise, a process for measuring the random noise value is added to the traditional measurement method. The specific process is as follows:

[0039] Step 1: In a dark environment, continuously capture images using an image sensor to obtain I-frames of image information with a resolution of M×N, where M is the number of pixels in the horizontal direction and N is the number of pixels in the vertical direction. Generally, the resolution of an image sensor is equal to the number of pixels in its pixel array. The grayscale values ​​of the image data form a matrix. Average the grayscale values ​​of each frame within the frame; this average value is recorded as the dark current value for each frame, where the average value of the i-th frame is expressed in μ... t_i This indicates that the value of i ranges from [1, I], and μ t_i Given a one-dimensional array with I elements

[0040]

[0041] In the formula, m and n represent the coordinates of different pixels in the pixel array of the image sensor, with m ranging from [0, M-1] and n ranging from [0, N-1]. t_i This represents the grayscale value of the pixel at coordinates (m,n) in the i-th frame.

[0042] Step 2: Calculate the standard deviation of the I dark current values, denoted as the dark current inter-frame fluctuation value BLC_stability:

[0043] BLC_stability=std(μ t_i (7)

[0044] Step 3: Average the grayscale value matrix of the I-frame image data obtained in Step 1 to obtain the average grayscale value matrix y. RN_ave .

[0045]

[0046] In the formula, m and n represent the coordinates of different pixels in the pixel array of the image sensor, with m ranging from [0, M-1] and n ranging from [0, N-1]. RN_i This represents the grayscale matrix of the i-th frame image data.

[0047] Step 4: From the I-frame image data obtained in Step 1, take the grayscale matrix y of any one frame of image data. RN_iSubtract the average gray value matrix y obtained in the third step from the gray value matrix of this frame image data. RN_ave After eliminating fixed-pattern noise, the resulting matrix is ​​a grayscale matrix containing only random noise. (Using y...) RN express.

[0048] y RN [m][n]=y RN_i [m][n]-y RN_ave [m][n] (9)

[0049] Step 5: Calculate the standard deviation of the grayscale matrix containing only random noise obtained in Step 4, denoted as random noise σ. RN (Grayscale value representation).

[0050] σ RN =std(y RN (10)

[0051] Step 6: Based on the expected value of the random noise residue in the aforementioned theory, the dark current inter-frame fluctuation value BLC_stability′ after removing the influence of random noise can be obtained as follows:

[0052]

[0053] Although the functions and working processes of the present invention have been described above in conjunction with the accompanying drawings, the present invention is not limited to the specific functions and working processes described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims, and all of these are within the protection scope of the present invention.

Claims

1. A method for measuring inter-frame fluctuations of dark current in an image sensor after removing the influence of random noise, characterized in that, Includes the following steps: 1) In a dark environment, the image sensor continuously captures images to obtain I-frame image information with a resolution of M×N, where M is the number of pixels in the horizontal direction of the image and N is the number of pixels in the vertical direction of the image; The grayscale values ​​of each frame of image data are averaged within the frame, and this average value is recorded as the dark current value of each frame, where the average value of the i-th frame is expressed in μ. t_i This indicates that the value of i ranges from [1, I], and μ t_i Given a one-dimensional array with I elements In the formula, m and n represent the coordinates of different pixels in the pixel array of the image sensor, with m ranging from [0, M-1] and n ranging from [0, N-1]. t_i This represents the grayscale value of the pixel at coordinates (m,n) in the i-th frame. 2) Calculate the standard deviation of I dark current values, denoted as the dark current inter-frame fluctuation value BLC_stability: BLC_stability=std(μ t_i ) 3) Average the grayscale value matrix of the I-frame image data obtained in step one to obtain the average grayscale value matrix y. RN_ave ; In the formula, m and n represent the coordinates of different pixels in the pixel array of the image sensor, with m ranging from [0, M-1] and n ranging from [0, N-1]. RN_i This represents the grayscale matrix of the i-th frame image data; 4) For the I-frame image data obtained in the first step, take the grayscale matrix y of any one frame of image data. RN_i Subtract the average gray value matrix y obtained in the third step from the gray value matrix of this frame image data. RN_ave The resulting matrix is ​​the gray-level matrix y containing only random noise. RN ; y RN [m][n]=y RN_i [m][n]-y RN_ave [m][n] 5) Calculate the standard deviation of the gray-level matrix containing only random noise obtained in step 4, denoted as random noise σ. RN ; s RN =std(y RN ) 6) Calculate the dark current inter-frame fluctuation value BLC_stability′ after removing the influence of random noise using the following formula.

Images