A refrigerated sCMOS scientific camera and its readout method

Abstract

This invention discloses a cooled sCMOS scientific camera and its readout method, relating to the fields of optical imaging and photoelectric detection technology. The camera includes a front housing, a rear housing, and a sealed cavity disposed between the front and rear housings. The sealed cavity is an airtight structure. An sCMOS image sensor is disposed inside the sealed cavity, near the incident light side. A heat sink, a thermoelectric cooler, and a second heat sink are sequentially disposed on the back of the image sensor. The cold side of the thermoelectric cooler is in contact with the first heat sink, and the hot side is connected to the second heat sink. A water-cooled plate is disposed at the end of the second heat sink away from the thermoelectric cooler. The other surface of the water-cooled plate is thermally connected to a main control circuit board via a third heat sink. The main control circuit board is located at the bottom of the rear housing. This invention effectively suppresses dark current noise by encapsulating the sensor in a vacuum-sealed cavity and introducing multi-stage cooling and heat dissipation technology.

Description

Technical Field

[0001] This invention relates to the fields of optical imaging and photoelectric detection technology, and in particular to a cooled sCMOS scientific camera and its readout method. Background Technology

[0002] In fields such as life sciences, astronomical observation, and semiconductor testing, image sensors often require long exposures under extremely low light conditions. Existing scientific cameras mainly suffer from the following problems: 1. High dark current noise: During long-term operation, the sensor chip heats up, causing the dark current to increase exponentially, which seriously pollutes the effective signal. Traditional air-cooling solutions have limited cooling capacity.

[0003] 2. High readout noise: In high-speed readout mode, the noise suppression capability of the front-end circuit is insufficient, resulting in a sharp decline in image quality at high frame rates.

[0004] 3. Stray light interference: The internal structure of the camera is not treated to eliminate light, and the incident light is reflected multiple times in the cavity, forming glare and reducing image contrast. Summary of the Invention

[0005] This invention provides a cooled sCMOS scientific camera, including a front camera housing 1, a rear camera housing 3, and a sealed cavity 2 disposed between the front camera housing 1 and the rear camera housing 3. The sealed cavity 2 is an airtight structure, and its interior is evacuated. An sCMOS image sensor 5 is disposed inside the sealed cavity 2, near the incident light side. A heat sink 7, a thermoelectric cooler 8, and a second heat sink 10 are sequentially disposed on the back of the image sensor 5. The cold side of the thermoelectric cooler 8 is in contact with the first heat sink 7, and the hot side is connected to the second heat sink 10. A water-cooled plate 11 is disposed at the end of the second heat sink 10 away from the thermoelectric cooler 8. The flow channel of the water-cooled plate 11 is a counter-flow channel, with the inlet and outlet of the water-cooled plate 11 respectively located at diagonal positions. The two water streams converge in the middle of the flow channel and then flow out in opposite directions. The other surface of the water-cooled plate 11 is thermally connected to a main control circuit board 13 through a third heat sink 12. The main control circuit board 13 is disposed at the bottom of the rear camera housing 3.

[0006] In the cooled sCMOS scientific camera described above, the sealed cavity 2 is composed of a front shell 21 and a rear shell 22, which are sealed together by a first sealing ring 16. The front shell 21 of the sealed cavity has a window hole for incident light to pass through, and the optical window 4 is pressed between the front shell 21 of the sealed cavity and the front shell 1 of the camera by a second sealing ring 17 to close the window hole.

[0007] In the cooled sCMOS scientific camera described above, the inner surface of the optical window 4 on the side facing the image sensor 5 is coated with a broadband anti-reflection film, and the outer surface on the side facing the incident light is deposited with an indium tin oxide transparent conductive shielding film.

[0008] In the cooled sCMOS scientific camera described above, indium sheets or highly thermally conductive graphene pads are filled between the image sensor 5 and the heat sink 7, between the heat sink 7 and the thermoelectric cooler 8, and between the thermoelectric cooler 8 and the heat sink 10.

[0009] As described above, in the cooled sCMOS scientific camera, vacuum feedthrough connectors 9 are symmetrically arranged on the rear shell 22 of the sealed cavity; an image sensor circuit board 6 electrically connected to the image sensor 5 is disposed inside the sealed cavity 2, and the image sensor circuit board 6 is connected to the vacuum feedthrough connector 9 via a flexible flat cable 14; the main control circuit board 13 is connected to the vacuum feedthrough connector 9 via a flexible flat cable 15.

[0010] In the cooled sCMOS scientific camera described above, the image sensor circuit board 6 is fixed to the rear shell 22 of the sealed cavity by a venting bolt 23 and a support column 24, and a wave spring is provided between the venting bolt 23 and the support column 24.

[0011] The cooled sCMOS scientific camera described above has a molecular sieve storage chamber 18 arranged inside the sealed cavity 2.

[0012] The cooled sCMOS scientific camera described above has a vacuum clamping tube 19 inserted through the rear shell 22 of the sealed cavity.

[0013] The present invention also provides a readout method for a cooled sCMOS scientific camera, comprising: Step S10: Real-time acquisition of the driving current of the semiconductor cooling chip and calculation of it into the real-time phase of water cooling pulsation; Step S20: Based on the calculated real-time phase statistics, read out the noise and calibrate the silent window; Step S30: Calculate and lock the readout trigger delay based on the real-time phase, and output the readout trigger signal aligned to the silent window; Step S40: Predict and fine-tune the exposure termination time so that the readout start time is aligned with the center phase of the silent window; Step S50: Correct image drift based on pulsating phase-displacement mapping table.

[0014] The beneficial effects achieved by this invention are as follows: By encapsulating the sensor in a vacuum-sealed cavity and introducing multi-stage cooling and heat dissipation technology, dark current noise is effectively suppressed; simultaneously, an integrated water-cooling structure is adopted to cool the heat-generating components of the entire device. The water-cooling structure uses a vortex-shaped flow channel, which reduces the disturbance force on the camera under the same thermal resistance. This enables control of pixel drift under long exposure conditions, increasing the stability of the camera. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0016] Figure 1 This is a structural appearance diagram of a cooled sCMOS scientific camera provided in Embodiment 1 of this application; Figure 2 This is a cross-sectional view of the overall structure of a cooled sCMOS scientific camera provided in Embodiment 1 of this application; Figure 3 yes Figure 2 Enlarged view of region A in the middle; Figure 4 This is an isometric view of the vacuum chamber of a cooled sCMOS scientific camera provided in Embodiment 1 of this application; Figure 5 This is a flow channel diagram of a water-cooled structure for a cooled sCMOS scientific camera provided in Embodiment 1 of this application. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Example 1

[0019] like Figure 1 As shown, this embodiment provides a cooled sCMOS scientific camera, including a front camera housing 1, a rear camera housing 3, and a sealed cavity 2 disposed between the front camera housing 1 and the rear camera housing 3.

[0020] refer to Figure 2 , Figure 3 The sealed cavity 2 is an airtight structure, and its interior is evacuated to a vacuum level not exceeding 1×10⁻⁶. -3 Pa is used to prevent frost formation inside the image sensor cavity under low-temperature cooling conditions. The sealed cavity 2 is composed of a front shell 21 and a rear shell 22. The front shell 21 and the rear shell 22 are sealed together by a first sealing ring 16 made of fluororubber. The front shell 21 has a window hole for incident light to pass through. The optical window 4 is pressed between the front shell 21 and the camera front shell 1 by a second sealing ring 17 (or a brazed seal can be added, i.e., the cavity and glass are transition welded with Kovar alloy) to close the window hole. Thus, the optical window 4, the front shell 21, and the rear shell 22 together form the sealed cavity 2.

[0021] An image sensor 5 is installed inside the sealed cavity 2 at the optical window 4. The image sensor 5 is a scientific-grade CMOS (sCMOS) image sensor used to capture optical signals incident through the optical window 4. An image sensor circuit board 6 is fixed on the rear shell 22 of the sealed cavity near the image sensor 5, and the image sensor 5 is electrically connected to the image sensor circuit board 6.

[0022] The inner surface of the optical window 4, on the side facing the image sensor 5, is coated with a broadband anti-reflection film, while the outer surface, on the side facing the incident light, is deposited with an indium tin oxide (ITO) transparent conductive shielding film to shield against electromagnetic interference.

[0023] A heat sink 7, made of oxygen-free copper, is located on the back of the image sensor 5. A thermoelectric cooler 8 is positioned at the end of the heat sink 7 furthest from the image sensor 5, with the cold surface of the heat sink 7 in contact with the cooler surface of the thermoelectric cooler 8. The thermoelectric cooler 8 has a maximum cooling capacity of no less than 50W, a maximum temperature difference of no less than 60℃, a rated operating voltage of 12V, a rated operating current of 6A, and operates in a constant temperature closed-loop control mode. To improve cooling efficiency, a 0.1mm thick indium sheet or high thermal conductivity graphene pad is filled between the image sensor 5 and the heat sink 7, and between the heat sink 7 and the thermoelectric cooler 8.

[0024] Vacuum feedthrough connectors 9 are symmetrically arranged on the rear shell 22 of the sealed cavity. A flexible flat cable socket is arranged on the image sensor circuit board 6. The vacuum feedthrough connectors 9 and the flexible flat cable socket on the image sensor circuit board 6 are connected by a flexible flat cable 14, thereby realizing the transmission of electrical signals between the image sensor 5 inside the sealed cavity 2 and the external circuit.

[0025] A heat sink 20 is installed in the middle of the rear shell 22 of the sealed cavity. The hot surface of the semiconductor cooling chip 8 is connected to the heat sink 20 through an indium sheet with a thickness of 0.1 mm or a high thermal conductivity graphene pad.

[0026] A water-cooled plate 11 is mounted on the end of heat sink 10 furthest from the thermoelectric cooler 8. The surfaces of heat sink 10 and water-cooled plate 11 are in contact with each other via thermally conductive silicone grease. Water-cooled plate 11 transfers heat to an external water chiller via water cooling. The other surface of water-cooled plate 11 is fixed to heat sink 12 via thermally conductive silicone grease. Heat sink 12 is used to dissipate heat for the main control circuit board 13 located at the bottom of the camera rear housing 3, thereby reducing the number of active / passive heat dissipation components such as fans, heat pipes, and vapor chambers, and reducing the overall size and complexity of the device.

[0027] The main control circuit board 13 is equipped with a flexible flat cable socket, which is connected to the vacuum feedthrough connector 9 via a flexible flat cable 15.

[0028] See Figure 4 The image sensor circuit board 6 is fixed to the rear shell 22 of the sealed cavity by the venting bolt 23 and the support column 24. A wave spring is provided between the venting bolt 23 and the support column 24 to absorb the deformation caused by thermal stress, thereby preventing the semiconductor cooling chip 8 from being damaged by thermal stress.

[0029] To maintain the vacuum level inside the sealed cavity 2 for as long as possible, a molecular sieve storage chamber 18 is also arranged in the sealed cavity 2 for storing molecular sieves, desiccants, or getters.

[0030] The rear shell 22 of the sealed cavity is provided with a vacuum clamping tube 19. The vacuum of the sealed cavity 2 is achieved by evacuating the vacuum through the vacuum clamping tube 19 and then clamping it to maintain a certain degree of vacuum.

[0031] See Figure 5 The water-cooled plate 11 has a flow channel design with opposing flow channels. The inlet and outlet are located at opposite corners of the water-cooled plate 11. The two water streams converge in the middle of the flow channel and then flow out in opposite directions. The purpose is to reduce the disturbance force of the cooling water on the entire water-cooled plate 11 and further reduce the vibration of the camera.

[0032] The heat dissipation path in this embodiment is as follows: heat generated by image sensor 5 → heat sink 1 7 → semiconductor cooling chip 8 → heat sink 2 10 → water cooling plate 11; heat generated by main control circuit board 13 → heat sink 3 12 → water cooling plate 11; and the heat is discharged to the outside of the sealed cavity 2 by water cooling plate 11 through forced water cooling.

[0033] The noise suppression method of this embodiment includes the following steps: after receiving an external trigger signal, when the temperature of the image sensor 5 is stabilized below a preset -20°C by the semiconductor cooling chip 8, the exposure timing is started; at the moment the exposure ends, the image processing circuit on the main control circuit board 13 samples the reset level and the signal level in sequence; and differential operations are performed on the reset level and the signal level to cancel the reset noise.

[0034] Example 2 Embodiment 2 of this application provides a readout method for a cooled sCMOS scientific camera. This method addresses the problem in Embodiment 1 where the water-cooled plate 11 experiences periodic micro-fluctuations in temperature due to water pump flow pulsation caused by the external water chiller, which are then conducted to the image sensor 5 via the heat sink 2 10 and the thermoelectric cooler 8, resulting in readout noise varying with the pulsation phase. The method indirectly obtains the water-cooled pulsation phase by detecting the driving current of the thermoelectric cooler 8 and locks the readout timing of the image sensor 5 to the phase window with the lowest pulsation noise, achieving low-noise imaging. The thermoelectric cooler 8 operates in a constant-temperature closed-loop control mode. The temperature change of the hot surface caused by the temperature fluctuation of the water-cooled plate 11 is converted into fluctuations in the driving current of the thermoelectric cooler 8 by the closed-loop control; this fluctuation component serves as the detection carrier for the water-cooled pulsation.

[0035] The overall data processing flow of this method is divided into a calibration stage and an operation stage. The calibration stage is performed once at the factory, sequentially completing the establishment of the phase calculation link in step S10, the silent window calibration in step S20, and the calibration of the phase-displacement mapping table in step S50. The operation stage is continuously executed during normal imaging. Step S10 continuously outputs real-time phase data, and steps S30 to S50 sequentially perform readout trigger locking, exposure termination adjustment, and image correction processing on the phase data.

[0036] For ease of description, the following parameters are defined: The water cooling pulsation frequency is estimated by step S10 and has a value range of 10Hz to 200Hz. This represents the water cooling pulsation period, with a value ranging from 5ms to 100ms. The phase of the water-cooling pulsation at time t is obtained by step S10; The center phase of the silent window is obtained by calibration in step S20; The phase range of the silent window is obtained by calibration in step S20; The reset level sampling time in the correlated double sampling With signal level sampling time The time interval between them is given by the datasheet of image sensor 5 or configured through the sensor timing register; in this implementation, it is 2 μs. The initial current sampling rate was set to 10kHz to cover the common water pump pulsation frequency range, and subsequently adjusted according to... Adjusted to no less than .

[0037] The method includes the following steps: Step S10: Real-time acquisition of the driving current of the semiconductor cooling chip and calculation of it into the real-time phase of water cooling pulsation; This step is performed in two stages: initial frequency estimation and real-time phase tracking. (I) Initial frequency estimation stage This phase is executed once when the system starts up or needs to be recalibrated. It uses a fixed sampling rate to collect data and estimates the initial value of the water-cooling pulsating frequency using FFT.

[0038] A 0.1Ω sampling resistor with a rated power of 5W is connected in series in the driving circuit of the semiconductor cooler 8. The differential voltage across the sampling resistor is input to a 12-bit analog-to-digital converter (ADC) at an initial sampling rate. Perform sampling conversion to obtain digital code. Since the driving current of the semiconductor cooling chip 8 is bidirectional, the analog-to-digital converter uses a differential input method to acquire positive and negative voltages. According to the code-to-voltage conversion relationship specified in the analog-to-digital converter datasheet, the... Converted to voltage across the sampling resistor ,Right now ,in The output code value of the ADC when there is zero input (given by the ADC datasheet, 2048 in this embodiment). This is the ADC reference voltage (given by the ADC datasheet; in this embodiment, it is taken as 2.5).

[0039] Then by Ohm's Law Obtain discrete sampled values ​​of the drive current ,in is the series resistance value, and n is the sampling point number.

[0040] First, read the discrete values ​​of the drive current from 8192 sampling points collected by the analog-to-digital converter. Calculate the average of these 8192 points. ,from Subtract To remove the DC component, obtain To suppress spectral leakage, The application of Hanning window The value obtained after windowing .

[0041] right The sequence is subjected to an 8192-point Fast Fourier Transform to obtain the frequency domain sequence. Take the amplitude spectrum Its frequency resolution is approximately 1.22 Hz. After eliminating the DC component (n=0), further power frequency interference points such as 50 Hz ± 2 Hz, 60 Hz ± 2 Hz, their second harmonics (100 Hz ± 4 Hz, 120 Hz ± 4 Hz), and third harmonics (150 Hz ± 6 Hz, 180 Hz ± 6 Hz) are eliminated. The frequency point corresponding to the maximum peak value in the remaining amplitude spectrum is denoted as... The initial estimated value of the water-cooled pulsating main frequency is: To eliminate misjudgments caused by occasional noise spikes, three consecutive FFT estimations are required. If the deviation of the three estimation results does not exceed ±5%, the average value is taken as the final initial main frequency; otherwise, the number of sampling points is increased to 16384 points and the estimation is repeated.

[0042] according to Adjust the ADC sampling rate to After reconfiguration, the data buffer is cleared. During reconfiguration, the phase output remains the phase value from the previous moment and is linearly extrapolated according to the previous tracking frequency. After reconfiguration, it switches to the real-time phase at the new sampling rate, entering the real-time phase tracking stage. All subsequent processing is based on the sampling rate. Execution, sampling period .

[0043] (II) Real-time phase tracking stage This phase runs continuously during normal imaging, and the following processing is performed sequentially on the current sampling point in each sampling cycle: With sampling rate Acquire ADC output code And convert it into the current drive current sampling value. ; Will By inputting a 4th-order Butterworth IIR digital bandpass filter, the AC component near the water-cooling pulsation frequency is obtained. The filter is implemented using a cascaded structure of four second-order sections, with a center frequency of [missing value]. (Initial run is) The upper and lower cutoff frequencies are respectively and coefficients of each second-order section , Based on the bilinear transform method and The coefficients are calculated using the cutoff frequency (e.g., output by MATLAB's `butter` function or Python's `scipy.signal.butter` function), quantized as 24-bit fixed-point numbers, and stored in the FPGA's internal coefficient ROM on the main control board 13. The difference equation for each second-order section is: ,in The input value for the current sampling point, This is the output value of the current sampling point. These are the input values ​​from the previous and the sample points, respectively. These are the output values ​​of the previous and the two preceding sampling points, respectively; the output after cascading four second-order sections is... .

[0044] The coefficient update frequency is no more than once every 100 sampling periods.

[0045] The FPGA's internal numerically controlled oscillator (NCO) on the main control circuit board 13 generates in-phase orthogonal discrete sequences cos[n] and sin[n] with the same frequency as the water-cooling pulsation. This NCO is implemented using a 32-bit phase accumulator and a sine lookup table (LUT) structure with a depth of 1024: the phase accumulator adds a frequency control word for each sampling period. The high 10 bits of the phase accumulator are used to address the LUT, and the output is 16-bit quantized. and .Will Multiply by the corresponding elements in cos[n] and sin[n] respectively to obtain the in-phase components after mixing. and orthogonal components .

[0046] Will and The baseband in-phase components are obtained by filtering out the second harmonic components using a first-order IIR low-pass filter. and orthogonal components The cutoff frequency of this low-pass filter is... Filter coefficients Since the mixed signal may contain higher harmonics introduced by NCO quantization and bandpass filtering non-ideals in addition to DC and second harmonics, the attenuation of the first-order low-pass filter is limited. Therefore, in this embodiment, the cutoff frequency of the low-pass filter is set to no more than 1 / 10 of the pulsation frequency to ensure that components higher than the second harmonic are attenuated to below -40dB.

[0047] Depend on The current pulsation phase is calculated. atan2 is implemented in an FPGA using a 16-stage pipelined CORDIC algorithm, with both input and output bit widths of 16 bits and an output range of [missing information]. .

[0048] Calculate the phase difference between adjacent sampling points ;like ,but ;like ,but ;otherwise This represents the phase after unwinding. To ensure reliable unwinding, the sampling rate must be no less than 5 times the pulsation frequency, at which point the phase change between adjacent sampling points should not exceed [the specified value]. Much smaller than the unwinding threshold Phase aliasing will not occur.

[0049] Instantaneous pulsation frequency calculated from the phase difference after unwinding , This represents the phase of the previous sampling point. To suppress noise, [the following is applied]... First-order smoothing is used to obtain the tracking frequency. ,in For the smoothing coefficient, take... =0.01, This is the tracking frequency that has been smoothed from the previous sampling point. Write to the NCO's frequency control word register and update. This allows the NCO output frequency to track the slow changes in the water-cooling pulsation frequency in real time.

[0050] The real-time phase after unwinding and tracking frequency The output is supplied to steps S20 to S50.

[0051] Step S20: Based on the calculated real-time phase statistics, read out the noise and calibrate the silent window; This step is performed once during the factory calibration phase, and the input is the real-time phase output from step S10. and The output is the center phase of the silent window. and phase range .

[0052] First, control the external water chiller to operate at the nominal flow rate (e.g., 1 L / min) to confirm that the phase calculation link in step S10 is working properly and outputting stably. and Place the image sensor 5 under light-blocking conditions (turn off the incident light or cover the lens cap), and configure the sensor to operate in the same gain, readout speed, and resolution modes as normal imaging to ensure that the calibration conditions are consistent with the actual usage conditions.

[0053] Then, within one complete cycle of water-cooling pulsation, at equal phase intervals... Set 36 target phases Using the readout phase-locking mechanism described in step S30, to... The delay is calculated for the target phase, and a readout trigger signal is output. M=100 offset frames are acquired. These offset frames are readout frames from the image sensor 5 under shading conditions. After each phase point is acquired, the actual phase value of that phase point is recorded. .

[0054] Next, the main control circuit board 13 performs statistical calculations on the acquired bias frames: for each actual phase The effective pixel area is defined as 80% of the pixels in the center of the photosensitive area of ​​image sensor 5. The temporal standard deviation of the grayscale value of 200 offset frames is calculated for each pixel position within the effective pixel area. The spatial average of the temporal standard deviations of all pixels within the effective pixel area is then taken as the readout noise for that phase. The basis for setting the number of frames acquired per phase point to 200 is that the statistical uncertainty of the standard deviation estimate is approximately... The time uncertainty is about 5%, and it can reliably distinguish noise differences of more than 10% between different phases; if higher calibration accuracy is required, it can be increased to 500 frames.

[0055] The threshold is set to 1.2 times the global minimum of the readout noise. ,Pick Less than The continuous phase interval is used as a silent window When multiple consecutive phase intervals satisfying the conditions exist, select the one containing... The continuous phase interval of the global minimum. Record the center phase of this silent window. Constraints are imposed on the width of the silent window: if the calibrated silent window width is less than 10°, it is expanded to 10° on both sides with the global minimum phase as the center to ensure sufficient phase margin during actual locking; if the silent window width is greater than 80° (i.e. the noise curve is too flat and phase locking is meaningless), an alarm message is recorded, prompting a check for abnormalities in the water cooling system or whether the sensor is operating at the normal cooling temperature, and the global minimum phase ±30° is taken as the silent window by default.

[0056] The quiescent window is defined relative to the phase of the water-cooled pulsation. When the flow rate of the external water chiller changes within ±50% of the nominal value, the frequency and amplitude of the water-cooled pulsation change, but the shape of the phase-noise relationship curve remains unchanged. Therefore, the center phase and width of the quiescent window do not need to be recalibrated. If the flow rate changes beyond ±50%, or after the water pump is replaced or the water-cooling system is maintained, this calibration step should be repeated.

[0057] Step S30: Calculate and lock the readout trigger delay based on the real-time phase, and output the readout trigger signal aligned to the silent window; This step continues during normal imaging, with the input being the real-time phase output from step S10. and and the output of step S20 The output is a readout trigger signal aligned to the silent window.

[0058] When the main control circuit board 13 receives an external exposure trigger signal or needs to initiate a frame readout, it determines the current real-time phase. Calculate the readout startup delay with the target phase: Where mod is the modulo operation, the result falls within the range of 0. Within the range. When the target phase exists, take Step S20 Calibration Stage Before output, the target phase is manually specified. .

[0059] The FPGA on the main control circuit board 13 is delayed by an internal high-precision timer (FPGA clock frequency is 100MHz, timing accuracy is 10ns). Then, a readout trigger signal is output from the external trigger input terminal of image sensor 5 to start the readout timing of image sensor 5, so that the reset level sampling time... Falling into the silent window Inside.

[0060] Due to the time interval of the related double sampling (In this example, it is 2μs) which is much smaller than the water-cooling pulsation period. (5ms to 100ms) Reset level sampling time With signal level sampling time phase difference between much smaller Therefore, as long as Falling into the silent window, It naturally falls into the silent window.

[0061] Step S40: Predict and fine-tune the exposure termination time so that the readout start time is aligned with the center phase of the silent window; This step is performed when the exposure timer is about to end, and the input is the real-time phase output from step S10. and tracking frequency and the output of step S20 .

[0062] Preset the exposure duration before the exposure timer ends. At that time, read the phase currently output in step S10. (corresponding time) and tracking frequency Assuming in Within the prediction window By keeping the water pump pulsation frequency constant (assuming this assumption holds), and performing a linear extrapolation of the phase, the preset exposure end time is predicted. phase ,in This is the preset exposure end time, given by the exposure timer, without any fine-tuning.

[0063] Subsequently, with the silent window center phase To achieve the target, calculate the fine-tuning amount. The result of the mod operation falls within Within the scope. Transform to a symmetric interval: If ,but ,like ,but Keeping it unchanged Finally landed Within the range, This refers to the water-cooling pulsation cycle. Regardless of... Whether or not the exposure is within the silent window, the exposure termination time is adjusted by the FPGA exposure timer on the main control circuit board 13. This aligns the readout start time with the center phase of the silent window. The exposure time deviation caused by fine-tuning shall not exceed ± (Typical value ±2.5ms to ±50ms), which is negligible for long exposures lasting seconds; for short exposure applications, this step can be turned off to avoid exposure time deviation.

[0064] Step S50: Correct image drift based on pulsating phase-displacement mapping table; This step is divided into two execution phases: calibration and execution. The calibration phase takes target images under different pulsation phases as input and outputs a phase-displacement mapping table. The input for the running phase is the phase time series output from step S10, denoted as... , where i is the sampling time index and the phase-displacement mapping table. The output is the drift-corrected image. This step assumes that the image drift caused by water-cooling pulsation is pure translation (without rotation or scaling). This assumption is based on the following: water-cooling pulsation causes periodic temperature fluctuations in the sensor mounting substrate and heat sink. Due to the mismatch in thermal expansion coefficients between the sensor chip and the mounting substrate, as well as the geometric asymmetry of the mounting structure, the temperature fluctuations cause the sensor as a whole to produce a small periodic shear displacement, which manifests as pure translation of the image. Rotation requires a non-uniform in-plane temperature gradient, and scaling requires the pixel size itself to change significantly with temperature (the temperature coefficient of sCMOS pixel size is usually less than 1ppm / ℃, and the scaling is less than 0.001% under ±1℃ fluctuation). In this scenario, both are much smaller in magnitude than translation (the typical translation amount is 0.01~0.1 pixels) and can be ignored.

[0065] (a) The execution process for the calibration phase is as follows: Using the readout phase-locking mechanism described in step S30, images of the random dot matrix test target are acquired at different phases of the water-cooling pulsation (e.g., one sampling point every 5°, for a total of 72 phase points). The image of the first sampling phase is used as the reference image. The remaining phase images are denoted as .

[0066] For each Compared with reference image Perform normalized cross-correlation calculations and take the position of the cross-correlation peak as the integer pixel displacement; perform one-dimensional parabolic fitting on the three adjacent correlation values ​​in the row and column where the peak is located to obtain a two-dimensional displacement vector with sub-pixel precision. ,in, For phase The amount of horizontal (column) translation of the lower image relative to the reference image (positive to the right). phase The vertical (row) translation of the lower image relative to the reference image (positive downwards). Establish a phase-displacement mapping table. ,in The data is stored in the non-volatile memory of the main control circuit board 13, and linear interpolation is used between adjacent sampling points.

[0067] (ii) The execution flow during the operation phase is as follows: For exposure time greater than a preset threshold A long exposure (twice the water-cooling pulsation period, i.e., not less than 10ms) is performed during the exposure at a sampling rate. Continuously record the pulsating phase time series output in step S10 (i = 1, 2, ..., N), where N is the number of sampling points during the exposure period; if the end time of the exposure is not an integer multiple of the sampling period, then the last sampling point takes the closest sampling time.

[0068] For each ,pass The instantaneous two-dimensional displacement vector at that moment can be obtained by looking up the table and using linear interpolation. ; Calculate the equivalent average displacement during the exposure period: This equivalent average displacement serves as a correction factor for long-exposure image drift.

[0069] For each pixel position (x, y) in the output image, the corresponding position in the original image Bilinear interpolation sampling is used to obtain the drift-corrected image. When the sampled coordinates exceed the boundaries of the original image, edge pixel copying is used to fill the gaps.

[0070] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made on the basis of the technical solution of the present invention should be included within the scope of protection of the present invention.

Claims

1. A cooled sCMOS scientific camera, characterized in that, It includes a front camera housing (1), a rear camera housing (3), and a sealed cavity (2) disposed between the front camera housing (1) and the rear camera housing (3); the sealed cavity (2) is an airtight structure and is evacuated inside. An sCMOS image sensor (5) is disposed inside the sealed cavity (2) on the side near the incident light. A heat sink (7), a thermoelectric cooler (8), and a heat sink (10) are disposed sequentially on the back of the image sensor (5). The cold side of the thermoelectric cooler (8) is in contact with the heat sink (7), and the hot side is connected to the heat sink (10). A water-cooled plate (11) is disposed at the end of the heat sink (10) away from the thermoelectric cooler (8). The flow channel of the water-cooled plate (11) is a counterflow channel. The inlet and outlet are respectively disposed at the diagonal positions of the water-cooled plate (11). The two water flows converge in the middle of the flow channel and then flow out in opposite directions. The other surface of the water-cooled plate (11) is thermally connected to the main control circuit board (13) through the heat sink (12). The main control circuit board (13) is disposed at the bottom of the camera back cover (3).

2. The cooled sCMOS scientific camera according to claim 1, characterized in that, The sealed cavity (2) is formed by combining the front shell (21) and the rear shell (22) of the sealed cavity, and the two are sealed by the first sealing ring (16); the front shell (21) of the sealed cavity is provided with a window hole for incident light to pass through, and the optical window (4) is pressed between the front shell (21) of the sealed cavity and the front shell (1) of the camera to close the window hole.

3. The cooled sCMOS scientific camera according to claim 2, characterized in that, The inner surface of the optical window (4) on the side facing the image sensor (5) is coated with a broadband anti-reflection film, and the outer surface on the side facing the incident light is deposited with an indium tin oxide transparent conductive shielding film.

4. The cooled sCMOS scientific camera according to claim 1, characterized in that, The image sensor (5) is filled with an indium sheet or a high thermal conductivity graphene pad between the image sensor (5) and the first heat sink (7), between the first heat sink (7) and the semiconductor cooling chip (8), and between the semiconductor cooling chip (8) and the second heat sink (10).

5. The cooled sCMOS scientific camera according to claim 1, characterized in that, Vacuum feedthrough connectors (9) are symmetrically arranged on the rear shell (22) of the sealed cavity; an image sensor circuit board (6) electrically connected to the image sensor (5) is provided inside the sealed cavity (2); the image sensor circuit board (6) is connected to the vacuum feedthrough connector (9) via a flexible flat cable (14); the main control circuit board (13) is connected to the vacuum feedthrough connector (9) via a flexible flat cable (15).

6. The cooled sCMOS scientific camera according to claim 5, characterized in that, The image sensor circuit board (6) is fixed to the rear shell (22) of the sealed cavity by a venting bolt (23) and a support column (24), and a wave spring is provided between the venting bolt (23) and the support column (24).

7. The cooled sCMOS scientific camera according to claim 1, characterized in that, The sealed cavity (2) is equipped with a molecular sieve storage chamber (18).

8. The cooled sCMOS scientific camera according to claim 2, characterized in that, A vacuum clamping tube (19) is installed on the rear shell (22) of the sealed cavity.

9. A readout method for a cooled sCMOS scientific camera, characterized in that, include: Step S10: Real-time acquisition of the driving current of the semiconductor cooling chip and calculation of it into the real-time phase of water cooling pulsation; Step S20: Based on the calculated real-time phase statistics, read out the noise and calibrate the silent window; Step S30: Calculate and lock the readout trigger delay based on the real-time phase, and output the readout trigger signal aligned to the silent window; Step S40: Predict and fine-tune the exposure termination time so that the readout start time is aligned with the center phase of the silent window; Step S50: Correct image drift based on pulsating phase-displacement mapping table.

Images