Backside illuminated CCD baseline drift correction method

Abstract

The invention discloses a backside illuminated CCD baseline drift correction method. The method comprises the following steps: acquiring dark noise of a plurality of horizontal pixels in a CCD to obtain pixel dark noise data; based on the pixel dark noise data, a fitting curved surface function is obtained through self-defined function fitting; substituting the temperature fed back by the temperature sensor and the set integral time into the fitting curved surface function, and calculating to obtain the minimum theoretical dark noise and the noise reduction amplitude of a plurality of horizontal pixels in the CCD; determining a target baseline, and judging whether the minimum theoretical dark noise is greater than the target baseline; and based on the noise reduction amplitude, filtering and noise reduction processing is carried out on the compensated dark noise, and correction is completed. The signal compensation circuit based on the subtracter is designed on hardware for signal compensation, filtering and noise reduction processing is carried out on software, the hardware and the software work cooperatively, the CCD baseline is effectively controlled to approach the target baseline, the fluctuation amplitude of the CCD baseline is reduced, CCD baseline drift correction is achieved, and the dynamic range of spectrograph detection is enlarged.

Description

Technical Field

[0001] The present invention relates to the field of back-illuminated CCD, and in particular to a back-illuminated CCD baseline drift correction method. Background Art

[0002] Spectral detection technology is widely used in fields such as biology, chemistry and polymer materials. Among them, back-illuminated CCD is used as a detector in most spectrometers due to its better quantum efficiency, large dynamic range, low dark noise and good response in the ultraviolet and near-infrared regions.

[0003] Back-illuminated CCDs will generate dark noises such as dark current noise and dark current shot noise during the detection process. The dark noise will increase with the increase of exposure time, i.e. integration time, and temperature, causing the CCD baseline to rise, i.e. serious drift. The fluctuation amplitude of the dark noise will also increase with the increase of CCD baseline drift, affecting the dynamic range of the spectrometer detection and the accuracy and reliability of the subsequent spectral analysis algorithm.

[0004] At present, back-illuminated CCDs are collected and controlled through correlated double sampling (CDS), but CDS has no obvious correction effect on the CCD baseline drift caused by dark noise related to time and temperature. In addition, a cooled CCD is used to control the dark current level and thus the dark noise, but the cost of a cooled CCD is relatively high. There is also a method of adding a refrigerant to the outside of the CCD to control the dark current level. The process of adding the refrigerant will cause a sudden drop in temperature, resulting in condensation on the photosensitive surface of the CCD, which will cause the spectrometer detection light to diverge severely and affect CCD detection.

[0005] Therefore, it is necessary to study how to effectively control dark noise and avoid serious drift of CCD baseline without using cooling CCD and various cooling operations. Summary of the invention

[0006] In order to solve the above-mentioned technical problems, the present invention provides a back-illuminated CCD baseline drift correction method, which realizes CCD baseline drift correction by designing a signal compensation circuit based on a subtractor in hardware to perform signal compensation and performing filtering and noise reduction processing in software, thereby increasing the dynamic range of spectrometer detection.

[0007] The technical solution adopted by the present invention is:

[0008] A back-illuminated CCD baseline drift correction method specifically comprises the following steps:

[0009] S1, collecting dark noise of multiple horizontal pixels inside the CCD to obtain pixel dark noise data;

[0010] S2, obtaining a fitting surface function by fitting a custom function based on pixel dark noise data;

[0011] S3, substituting the temperature fed back by the temperature sensor and the set integration time into the fitting surface function, and calculating the minimum theoretical dark noise and the noise reduction amplitude of multiple horizontal pixels inside the CCD;

[0012] S4, determine the target baseline, and judge whether the minimum theoretical dark noise is greater than the target baseline: if so, perform signal compensation on the CCD; otherwise, perform filtering and noise reduction on the dark noise after normal acquisition;

[0013] S5. Based on the noise reduction amplitude, filter and reduce noise on the compensated dark noise to complete correction.

[0014] Furthermore, the pixel dark noise data in step S1 includes first pixel dark noise data and second pixel dark noise data; wherein the first pixel dark noise data is obtained by controlling the integration time and collecting the dark noise of multiple horizontal pixels inside the CCD at different temperatures; the second pixel dark noise data is obtained by controlling the temperature and collecting the dark noise of multiple horizontal pixels inside the CCD at different integration times.

[0015] Furthermore, the step S2 specifically includes the following steps:

[0016] 1) Based on the same horizontal pixel, the relationship between horizontal pixel dark noise and integration time is obtained by least squares linear fitting. The specific formula is as follows:

[0017] N p (t) = a p (T)t+b p (T),p=1,2,3...P(1)

[0018] In formula (1), N p is the horizontal pixel dark noise; p is the horizontal pixel; P is the horizontal pixel value; t is the integration time; T is the temperature; a p , b p are the fitting straight line coefficients corresponding to the horizontal pixel p respectively;

[0019] 2) Based on the same horizontal pixel, the relationship between horizontal pixel dark noise and temperature is obtained by curve fitting. The specific formula is as follows:

[0020]

[0021] In formula (2), N p is the horizontal pixel dark noise; p is the horizontal pixel; P is the horizontal pixel value; t is the integration time; T is the temperature; c p d pare the fitting curve coefficients corresponding to the horizontal pixel p respectively;

[0022] 3) Based on the same horizontal pixel, the relationship between the horizontal pixel dark noise, integration time and temperature is obtained by surface fitting, that is, the fitting surface function. The specific formula is as follows:

[0023]

[0024] In formula (3), N p is the horizontal pixel dark noise; p is the horizontal pixel; P is the horizontal pixel value; t is the integration time; T is the temperature; a p , b p 、c p d p are the fitting coefficients corresponding to the horizontal pixel p; f p is the corresponding horizontal pixel dark noise N p The correction constant.

[0025] Furthermore, the step S4 specifically includes the following steps:

[0026] 1) Determine the target baseline: According to the commonly used CCD dark noise standard, when the CCD baseline is controlled at around 2000, the maximum signal-to-noise ratio is as high as 33dB, indicating that the dark noise level at this time is already close to zero, so the target baseline is set to 2000;

[0027] 2) Determine whether the minimum theoretical dark noise is greater than the target baseline of 2000: If so, it indicates that the CCD baseline has drifted and is too high, and the CCD signal is compensated to make the CCD baseline close to the target baseline; otherwise, it indicates that the CCD baseline has not drifted, and the dark noise is filtered and denoised after normal acquisition;

[0028] Among them, signal compensation for CCD includes designing a signal compensation circuit and calculating a signal compensation voltage;

[0029] ① Design process of signal compensation circuit: set up signal compensation circuit between CCD and A / D converter, specifically: after CCD completes one acquisition, the pixel dark noise data output is sent to A / D converter through signal compensation circuit; then A / D converter sends the converted data to FPGA chip through synchronization; finally, it is sent to PC via USB.

[0030] ②The process of calculating signal compensation voltage:

[0031] The minimum theoretical dark noise in one acquisition is obtained by fitting the surface function, and the compensation dark noise in the signal compensation circuit is obtained based on the target baseline of the CCD dark noise. The specific calculation formula is as follows:

[0032] ΔN min =Nmin -N Bl (8)

[0033] In formula (8), ΔN min is the compensation dark noise in the signal compensation circuit; N min is the minimum theoretical dark noise; N Bl is the target baseline of CCD dark noise;

[0034] The internal reference voltage of the A / D converter is obtained based on the 16-bit precision A / D converter AD9826;

[0035] The signal compensation voltage V is calculated based on the compensated dark noise and the internal reference voltage of the A / D converter. com , the specific calculation formula is as follows:

[0036]

[0037] In formula (9), V com is the signal compensation voltage; ΔN min is the compensation dark noise in the signal compensation circuit; V ref The internal reference voltage of the A / D converter is 4V.

[0038] Furthermore, the step S5 specifically includes the following steps:

[0039] The original dark noise N of the signal compensation circuit 1p Multiple horizontal pixels are compensated to compensate for dark noise ΔN min The value of the new dark noise N1' generated after denoising p , the specific calculation formula is as follows:

[0040] N1' p =N 1p -(N 2p -ΔN min )(10)

[0041] In formula (10), N1' p is the new dark noise generated after denoising; N 1p is the original dark noise after the signal compensation circuit; N 2p The dark noise under the conditions of integration time and temperature is simulated by fitting surface function; ΔN min Compensate for dark noise in the signal compensation circuit.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] The present invention designs a signal compensation circuit based on a subtractor in hardware to perform signal compensation and performs filtering and noise reduction processing in software. The hardware and software work together to effectively control the CCD baseline to approach the target baseline, reduce the fluctuation amplitude of dark noise, realize CCD baseline drift correction, increase the dynamic range of spectrometer detection, and ensure the accuracy and reliability of subsequent spectral analysis algorithms.

[0044] 2) Control dark noise without cooling CCD and various cooling operation conditions, reduce costs and ensure normal CCD detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments:

[0046] Figure 1 It is a flow chart of a back-illuminated CCD baseline drift correction method of the present invention;

[0047] Figure 2 is a schematic diagram of a fitted straight line between horizontal pixel dark noise and integration time in the present invention;

[0048] Figure 3 is a schematic diagram of a fitting curve of horizontal pixel dark noise and temperature in the present invention;

[0049] Figure 4 It is a fitted three-dimensional surface diagram of horizontal pixel dark noise, integration time and temperature in the present invention;

[0050] Figure 5 It is a logic judgment schematic diagram of a back-illuminated CCD baseline drift correction method of the present invention;

[0051] Figure 6 It is a schematic diagram of the system framework design with a signal compensation circuit in the present invention;

[0052] Figure 7 It is a schematic diagram of the CCD output signal voltage detected by the oscilloscope in the present invention;

[0053] Figure 8 It is a schematic diagram of the signal voltage change when the CCD baseline drifts in the present invention;

[0054] Fig. 9 It is a schematic diagram of the working principle of the signal compensation circuit in the present invention;

[0055] Fig.10 It is a schematic diagram of experimental results using the correction method of the present invention. DETAILED DESCRIPTION

[0056] Example

[0057] like Figures 1 to 10As shown, a back-illuminated CCD baseline drift correction method specifically includes the following steps:

[0058] S1, collecting dark noise of multiple horizontal pixels inside the CCD to obtain pixel dark noise data;

[0059] Specifically, the pixel dark noise data in step S1 includes first pixel dark noise data and second pixel dark noise data; wherein the first pixel dark noise data is obtained by controlling the integration time and collecting the dark noise of multiple horizontal pixels inside the CCD at different temperatures; the second pixel dark noise data is obtained by controlling the temperature and collecting the dark noise of multiple horizontal pixels inside the CCD at different integration times;

[0060] In this embodiment, the number of horizontal pixels p is set to 2048.

[0061] S2, obtaining a fitting surface function by fitting a custom function based on pixel dark noise data;

[0062] Specifically, 1) based on the same horizontal pixel, the relationship between the horizontal pixel dark noise and the integration time is obtained by least squares linear fitting, that is, the fitting straight line function, and the specific formula is as follows:

[0063] N p (t) = a p (T)t+b p (T),p=1,2,3...P(1)

[0064] In formula (1), N p is the horizontal pixel dark noise; p is the horizontal pixel; P is the horizontal pixel value; t is the integration time; T is the temperature; a p , b p are the fitting straight line coefficients corresponding to the horizontal pixel p respectively;

[0065] In this embodiment, Figure 2 As shown, Figure 2 (a), (b), and (c) respectively represent the horizontal pixel dark noise N when the horizontal pixel p is 500, 1000, and 1500. p The fitting straight line with the integration time t, where the abscissa represents the integration time t and the ordinate represents the horizontal pixel dark noise N p , red, blue, and black represent temperatures T of 8°C, 18°C, and 28°C, respectively; Figure 3 It can be seen that the horizontal pixel dark noise N p It is linearly related to the integration time t, and the slope of the fitting line gradually increases with the increase of temperature T; Figure 2 (a), (b), and (c) in the figure show that the inherent properties of each horizontal pixel of the CCD make the fitting lines at different horizontal pixel positions have different coefficients;

[0066] 2) Based on the same horizontal pixel, the horizontal pixel dark noise N is obtained by curve fitting p The relationship with temperature T is the fitting curve function, and the specific formula is as follows:

[0067]

[0068] In formula (2), N p is the horizontal pixel dark noise; p is the horizontal pixel; P is the horizontal pixel value; t is the integration time; T is the temperature; c p d p are the fitting curve coefficients corresponding to the horizontal pixel p respectively;

[0069] In this embodiment, Figure 3 As shown, Figure 3 (a), (b), and (c) respectively represent the horizontal pixel dark noise N when the horizontal pixel p is 500, 1000, and 1500. p The fitting curve with temperature T, where the abscissa represents temperature T and the ordinate represents horizontal pixel dark noise N p , red, blue and black represent the integration time t of 10s, 30s and 60s respectively; Figure 4 It can be seen that the horizontal pixel dark noise N p It has an exponential relationship with the temperature T, and the curvature of the fitting curve gradually increases with the increase of the integration time t; in addition Figure 3 The inherent properties of each horizontal pixel of the CCD in (a), (b), and (c) make the fitting curves at different horizontal pixel positions have different coefficients;

[0070] 3) Based on the same horizontal pixel, the relationship between the horizontal pixel dark noise, integration time and temperature is obtained by surface fitting, that is, the fitting surface function. The specific formula is as follows:

[0071]

[0072] In formula (3), N p is the horizontal pixel dark noise; p is the horizontal pixel; P is the horizontal pixel value; t is the integration time; T is the temperature; a p , b p 、c p d p are the fitting coefficients corresponding to the horizontal pixel p; f p is the corresponding horizontal pixel dark noise N p The correction constant of

[0073] In this embodiment, Figure 4The figure shows the fitted three-dimensional surface diagram of horizontal pixel dark noise, integration time and temperature. The specific fitting process is as follows: taking horizontal pixel p=500 as an example, firstly, using formula (1) to obtain the fitted straight line function in the interval of t=8ms-60s, then arranging the fitted straight line function in the range of T=0℃-28℃ to obtain a three-dimensional matrix, and finally performing surface fitting to obtain a three-dimensional schematic diagram of the fitted surface function.

[0074] In addition, in the fitting process, the coefficient of determination R 2 (0≤R 2 ≤1) The fitting effect of the fitting straight line function, the fitting curve function and the fitting surface function was evaluated in turn. Determination coefficient R 2 The larger it is and the closer it is to 1, the better the fitting effect is. In this example, through Matlab calculation, Figure 2-4 The determination coefficients of the fitted straight line function, the fitted curve function and the fitted surface function based on the same horizontal pixel all reached 0.99≤R 2 ≤1, indicating that the fitting straight line function, fitting curve function and fitting surface function are consistent with the actual situation.

[0075] S3, substituting the temperature feedback from the temperature sensor and the set integration time into the fitting surface function, and calculating the minimum theoretical dark noise and the noise reduction amplitude of multiple horizontal pixels inside the CCD;

[0076] S4, determine the target baseline, and judge whether the minimum theoretical dark noise is greater than the target baseline: if so, perform signal compensation on the CCD; otherwise, perform filtering and noise reduction on the dark noise after normal acquisition;

[0077] Specifically, 1) determine the target baseline: according to the commonly used CCD dark noise standard, when the CCD baseline is controlled at about 2000, the maximum signal-to-noise ratio is as high as 33dB, indicating that the dark noise level at this time is already close to zero, so the target baseline is set to 2000;

[0078] 2) If Figure 5 As shown, it is determined whether the minimum theoretical dark noise is greater than the target baseline 2000: if so, it indicates that the CCD baseline has drifted and the baseline is too high, and the CCD signal is compensated to make the CCD baseline close to the target baseline; otherwise, it indicates that the CCD baseline has not drifted, and the dark noise is filtered and denoised after normal acquisition;

[0079] Among them, signal compensation for CCD includes designing a signal compensation circuit and calculating a signal compensation voltage;

[0080] ① Design process of signal compensation circuit: Figure 6As shown, a signal compensation circuit is set between the CCD and the A / D converter. Specifically, the pixel dark noise data output by the CCD after completing one acquisition is sent to the A / D converter through the signal compensation circuit; then the A / D converter sends the converted data to the FPGA on-chip RAM through synchronization; finally, it is sent to the PC via USB2.0.

[0081] The working principle of the signal compensation circuit: Fig. 9 As shown, the reset signal is compensated based on the voltage change relationship between the horizontal pixel signal and the reset signal, as follows: 1) When the CCD analog signal, i.e., the CCD dark noise signal, enters the reset signal stable stage, the switch S1 is connected to the D / A converter, the switch S2 is connected to the resistor R4, and the system forms a subtractor circuit. The instruction data calculated by the fitting function in the library is transmitted through I 2 The C serial bus controls the D / A converter to output a signal compensation voltage, and the reset signal is compensated by a differential circuit subtractor composed of an operational amplifier; 2) After the reset signal is acquired, the CCD analog signal enters the horizontal pixel signal output stage, and switches S1 and S2 are immediately disconnected, and the system forms a follower circuit to acquire the effective signal: if the minimum theoretical dark noise calculated by the fitting surface function is less than the target baseline, signal compensation will not be performed in this acquisition, and switches S1 and S2 remain disconnected to form a follower circuit.

[0082] The principle of signal compensation voltage is as follows: Figure 7 As shown, it represents the CCD output signal voltage detected by the oscilloscope, wherein label ① represents the reset buffer stage, label ② represents the reset signal stabilization stage, and label ③ represents the horizontal pixel signal output stage.

[0083] In this embodiment, the correlated double sampling method is used to collect the reset signal and the horizontal pixel signal voltage and eliminate the reset signal voltage by subtracting the horizontal pixel signal voltage from the reset signal to complete the signal extraction. Assuming that the dark noise is 0, the CCD output signal voltage calculation formula after correlated double sampling is as follows:

[0084] V out =V res -V signal (4)

[0085] In formula (4), V out is the CCD output signal voltage after correlated double sampling; V res is the reset signal voltage; V signal is the horizontal pixel signal voltage;

[0086] In this embodiment, long-term integration will affect the noise level of the effective signal. As the integration time increases, the charge accumulation inside the CCD will also increase, and the increase in temperature will also accelerate the rate of charge accumulation. These factors will cause the horizontal pixel signal voltage to decrease, that is, V signal The value decreases, plus the reset signal voltage V res It will not change with the change of integration time, which will cause the CCD output signal voltage to increase, which is manifested in the graphical interface as the CCD baseline rising and the noise floating level also increasing.

[0087] like Figure 8 The figure shows the schematic diagram of the signal voltage change when the CCD baseline drifts, where V s ' ignal >V Bl , the specific formula is as follows:

[0088] V n =V Bl -V s ' ignal (5)

[0089] In formula (5), V n is the charge accumulation voltage when the CCD baseline drifts, that is, V Bl With V s ' ignal The pressure difference; V Bl is the target baseline N Bl The corresponding horizontal pixel signal voltage; V s ' ignal is the horizontal pixel signal voltage when the CCD baseline drifts;

[0090] When the integration time and temperature change and charge accumulation occurs, the CCD reset signal voltage V res Will not change accordingly, at this time the CCD output signal voltage V out The output signal voltage after baseline drift is as follows:

[0091] V out =V res -(V Bl -V n ) (6)

[0092] In formula (6), V out is the CCD output signal voltage after correlated double sampling; V res is the reset signal voltage; V Bl is the target baseline N Bl The corresponding horizontal pixel signal voltage; V n is the charge accumulation voltage when the CCD baseline drifts;

[0093] In addition, the reset signal voltage V of the CCD needs to be adjusted before the correlated double sampling. res Compensation is performed. At this time, the CCD output signal voltage calculation formula after baseline correction is as follows:

[0094] V out =(V res -V com )-(V Bl -V n ) (7)

[0095] In formula (7), V out is the CCD output signal voltage after correlated double sampling; V res is the reset signal voltage; V com is the signal compensation voltage; V Bl is the target baseline N Bl The corresponding horizontal pixel signal voltage; V n is the charge accumulation voltage when the CCD baseline drifts;

[0096] When V com =V n , can reduce V out float, thereby correcting the CCD dark noise baseline drift.

[0097] ②The process of calculating signal compensation voltage:

[0098] The minimum theoretical dark noise in one acquisition is obtained by fitting the surface function, and the compensation dark noise in the signal compensation circuit is obtained based on the target baseline of the CCD dark noise. The specific calculation formula is as follows:

[0099] ΔN min =N min -N Bl (8)

[0100] In formula (8), ΔN min is the compensation dark noise in the signal compensation circuit; N min is the minimum theoretical dark noise; N Bl is the target baseline of CCD dark noise;

[0101] The internal reference voltage of the A / D converter is obtained based on the 16-bit precision A / D converter AD9826;

[0102] The signal compensation voltage V is calculated based on the compensated dark noise and the internal reference voltage of the A / D converter. com , the specific calculation formula is as follows:

[0103]

[0104] In formula (9), V comis the signal compensation voltage; ΔN min is the compensation dark noise in the signal compensation circuit; V ref The internal reference voltage of the A / D converter is 4V.

[0105] In this embodiment, the minimum theoretical dark noise N min As a standard, the dark noise compensation ΔN is obtained min , which can effectively avoid the problem of weakening the effective signal due to excessive compensation of dark noise; for different models of CCD, acquisition fitting operations at different integration times and temperatures are performed before installation, and multiple fitting functions obtained by fitting are built into a library and placed in the spectrometer. When the spectrometer is working, input the operation instructions including the integration time, and use the temperature sensor to feedback the temperature in real time, and then call the corresponding fitting function in the library to obtain the signal compensation voltage.

[0106] S5, based on the noise reduction amplitude, filtering and noise reduction processing is performed on the compensated dark noise to complete the correction;

[0107] Specifically, the original dark noise N of the signal compensation circuit 1p The 2048 horizontal pixels are compensated for the dark noise ΔN min The value of the new dark noise N1' generated after denoising p , the specific calculation formula is as follows:

[0108] N1' p =N 1p -(N 2p -ΔN min )(10)

[0109] In formula (10), N1' p is the new dark noise generated after denoising; N 1p is the original dark noise after the signal compensation circuit; N 2p The dark noise under the conditions of integration time and temperature is simulated by fitting surface function; ΔN min To compensate for dark noise in a signal compensation circuit;

[0110] In this embodiment, the dark noise values ​​of 2048 horizontal pixels inside the CCD have been compensated by ΔN through the signal compensation circuit. min , so that the CCD baseline drift range is controlled to a certain extent, but the fluctuation amplitude of CCD dark noise is still large at this time, and it is necessary to reduce the fluctuation amplitude of dark noise through noise reduction means; then through 2048 fitting surface functions, the dark noise N under the conditions of integration time t and temperature T is simulated 2p (p=1,2,3...P), and compare it with the original dark noise N after the signal compensation circuit and collected by A / D9826 1p(p=1,2,3...P) performs subtraction operation to eliminate the influence of random errors to the greatest extent; therefore, after filtering and noise reduction processing, the dark noise fluctuation amplitude of all horizontal pixels of CCD is accurately reduced without affecting the effective signal, further ensuring the stability of the dark noise baseline near the target baseline;

[0111] The original signal of the CCD is subjected to signal compensation and filtering noise reduction processing using the correction method of the present invention, and then the effect of the CCD baseline drift correction is analyzed. Fig.10 As shown in Figure 1, it shows the change of CCD baseline before and after correction at different integration times and temperatures. First, three groups of different integration times were tested at room temperature T = 28°C, as shown in Figure 1. Fig.10 (a), (b), and (c); secondly, at t = 60s, three groups of different temperatures were tested. Fig.10 As shown in (d), (e), and (f), the experimental results are quite significant.

[0112] In order to verify the effect of the correction method of the present invention, the effect of CCD baseline correction is evaluated by the average value AVG and the effective value RMSE of the CCD dark noise, wherein the AVG and RMSE of the CCD dark noise output data respectively reflect the numerical value of the CCD baseline and the strength of the noise volatility. In order to achieve the stability control of the CCD baseline drift and the near-zero baseline level control, the AVG and RMSE of the CCD dark noise output data are as small as possible. In this embodiment, the CCD integration time is between 8ms and 60s, and the temperature is within 0℃-28℃. Therefore, it is necessary to ensure that the CCD baseline is stable and as close to zero as possible within the integration time and temperature range. The specific results are shown in Table 1:

[0113] Table 1 Comparison results of CCD output signal voltage before and after correction

[0114]

[0115] It can be seen from Table 1 that when t=60s, T=28°C, the CCD baseline is about 29500, indicating that the degree of CCD baseline drift is extremely serious; after adopting the method of this embodiment, the CCD baseline is controlled at about the target baseline of 2000, and the dark noise baseline correction percentage is as high as 93.41%, that is, within the integration time range of 8ms-60s and the temperature range of 0°C-28°C, the CCD baseline is stable at about the target baseline of 2000, which has met the experimental requirements and meets the dark noise baseline stability requirements of most high-performance spectrometers.

[0116] In addition, after the CCD dark noise is reduced, the fluctuation amplitude of the CCD dark noise is also extremely well controlled. Under the condition of T = 28 ° C, when the maximum integration time is t = 60s, the RSME is only about 570.0, and the fluctuation amplitude of the CCD dark noise is small. For a spectrometer with a maximum CCD dynamic range of 65535, the maximum signal-to-noise ratio can still be as high as 28db when the integration time is 60s.

[0117] The present invention designs a signal compensation circuit based on a subtractor in hardware to perform signal compensation on the analog signal output by the CCD, and performs filtering and noise reduction processing on the digital signal collected by the A / D converter in software. The hardware and software work together to effectively control the CCD baseline to approach the target baseline, reduce the fluctuation amplitude of the CCD baseline, realize CCD baseline drift correction and near-zero baseline level control, increase the dynamic range of spectrometer detection, and ensure the accuracy and reliability of subsequent spectral analysis algorithms. In addition, dark noise is controlled without the need for a refrigerated CCD and various refrigeration operations, thereby reducing costs and ensuring normal CCD detection.

[0118] The embodiments described above are merely descriptions of preferred implementation modes of the present invention and are not intended to limit the scope of the present invention. Without departing from the principles and essence of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the scope of protection determined by the claims of the present invention.

Claims

1. A back-illuminated CCD baseline drift correction method, characterized in that: The specific steps include: S1, collecting dark noise of multiple horizontal pixels inside the CCD to obtain pixel dark noise data; S2, obtaining a fitting surface function by fitting a custom function based on pixel dark noise data; S3, substituting the temperature fed back by the temperature sensor and the set integration time into the fitting surface function, and calculating the minimum theoretical dark noise and the noise reduction amplitude of multiple horizontal pixels inside the CCD; S4, determine the target baseline, and judge whether the minimum theoretical dark noise is greater than the target baseline: if so, perform signal compensation on the CCD; Otherwise, the dark noise is filtered and denoised after normal acquisition; S5. Based on the noise reduction amplitude, filter and reduce noise on the compensated dark noise to complete correction.

2. A back-illuminated CCD baseline drift correction method according to claim 1, characterized in that: The pixel dark noise data in step S1 includes first pixel dark noise data and second pixel dark noise data; wherein the first pixel dark noise data is obtained by controlling the integration time and collecting the dark noise of multiple horizontal pixels inside the CCD at different temperatures; the second pixel dark noise data is obtained by controlling the temperature and collecting the dark noise of multiple horizontal pixels inside the CCD at different integration times.

3. The method for correcting baseline drift of a back-illuminated CCD according to claim 1, characterized in that: The step S2 specifically includes the following steps: 1) Based on the same horizontal pixel, the relationship between horizontal pixel dark noise and integration time is obtained by least squares linear fitting. The specific formula is as follows: N p (t)=a p (T)t+b p (T),p=1,2,3...P(1) In formula (1), N p is the horizontal pixel dark noise; p is the horizontal pixel; P is the horizontal pixel value; t is the integration time; T is temperature; a p 、b p are the fitting straight line coefficients corresponding to the horizontal pixel p respectively; 2) Based on the same horizontal pixel, the relationship between horizontal pixel dark noise and temperature is obtained by curve fitting. The specific formula is as follows: In formula (2), N p is the horizontal pixel dark noise; p is the horizontal pixel; P is the horizontal pixel value; t is the integration time; T is temperature; c p d p are the fitting curve coefficients corresponding to the horizontal pixel p respectively; 3) Based on the same horizontal pixel, the relationship between the horizontal pixel dark noise, integration time and temperature is obtained by surface fitting, that is, the fitting surface function. The specific formula is as follows: In formula (3), N p is the horizontal pixel dark noise; p is the horizontal pixel; P is the horizontal pixel value; t is the integration time; T is temperature; a p 、b p 、c p d p are the fitting coefficients corresponding to the horizontal pixel p; f p is the corresponding horizontal pixel dark noise N p The correction constant.

4. The method for correcting baseline drift of a back-illuminated CCD according to claim 1, characterized in that: The step S4 specifically includes the following steps: 1) Determine the target baseline: According to the commonly used CCD dark noise standard, when the CCD baseline is controlled at around 2000, the maximum signal-to-noise ratio is as high as 33dB, indicating that the dark noise level at this time is already close to zero, so the target baseline is set to 2000; 2) Determine whether the minimum theoretical dark noise is greater than the target baseline of 2000: If so, it indicates that the CCD baseline has drifted and is too high. Perform signal compensation on the CCD to make the CCD baseline approach the target baseline; Otherwise, it indicates that the CCD baseline has not drifted, and the dark noise is filtered and de-noised after normal acquisition; Among them, signal compensation for CCD includes designing a signal compensation circuit and calculating a signal compensation voltage; ① Design process of signal compensation circuit: set up signal compensation circuit between CCD and A / D converter, specifically: pixel dark noise data generated by CCD after completing one acquisition is sent to A / D converter through signal compensation circuit; then A / D converter sends the converted data to FPGA chip through synchronization; finally, it is sent to PC via USB. ②The process of calculating signal compensation voltage: The minimum theoretical dark noise in one acquisition is obtained by fitting the surface function, and the compensation dark noise in the signal compensation circuit is obtained based on the target baseline of the CCD dark noise. The specific calculation formula is as follows: ΔN min =N min -N Bl (8) In formula (8), ΔN min is the compensation dark noise in the signal compensation circuit; N min is the minimum theoretical dark noise; N Bl is the target baseline of CCD dark noise; The internal reference voltage of the A / D converter is obtained based on the 16-bit precision A / D converter AD9826; The signal compensation voltage V is calculated based on the compensated dark noise and the internal reference voltage of the A / D converter. com , the specific calculation formula is as follows: In formula (9), V com is the signal compensation voltage; ΔN min is the compensation dark noise in the signal compensation circuit; V ref The internal reference voltage of the A / D converter is 4V.

5. The method for correcting baseline drift of a back-illuminated CCD according to claim 1, characterized in that: The step S5 specifically includes the following steps: The original dark noise N of the signal compensation circuit 1p Multiple horizontal pixels are compensated to compensate for dark noise ΔN min The value of the new dark noise N1' generated after denoising p , the specific calculation formula is as follows: N1' p =N 1p -(N 2p -ΔN min )(10) In formula (10), N1' p is the new dark noise generated after denoising; N 1p is the original dark noise after the signal compensation circuit; N 2p The dark noise under the conditions of integration time and temperature is simulated by fitting surface function; ΔN min Compensate for dark noise in the signal compensation circuit.

Images