Non-uniformity correction method and device for point-by-point time domain calibration of infrared detector

Abstract

The invention discloses a non-uniformity correction method and device for point-by-point time domain calibration of an infrared detector. The non-uniform correction method for point-by-point time domain calibration of the infrared detector comprises the following steps of: irradiating each pixel in the infrared detector point by point by adopting a preset pulse black body source as standard radiation so as to obtain simulation response data of each pixel in the infrared detector; according to the response data of each pixel in the infrared detector, constructing a simulated thermal response transfer function corresponding to each pixel; and performing non-uniformity correction on the infrared detector according to the simulated thermal response transfer function corresponding to each pixel. The simulated thermal response transfer function of each pixel can reflect the corresponding dynamic response mechanism, the simulated thermal response transfer function is corrected through the dynamic response mechanism, the dynamic application scene of infrared imaging is met, full-working-condition pixel-level response consistency remodeling is achieved, bottom support is provided for high-frame-frequency and wide-temperature-range infrared imaging, and the method is suitable for large-scale popularization and application. The method is suitable for mass production calibration of large-area-array detectors.

Description

Technical Field

[0001] This invention relates to the field of infrared detector technology, and more specifically, to a method and apparatus for non-uniformity correction of point-by-point time-domain calibration of infrared detectors. Background Technology

[0002] The non-uniformity of the spatial distribution of infrared images can severely affect the performance of infrared imaging systems and adversely impact their use. Therefore, non-uniformity correction is necessary to meet practical application requirements. The non-uniformity of infrared imaging systems mainly stems from the non-uniformity of the infrared detector response and the non-uniformity of the readout circuit electronics. Among these, the temperature response of the infrared detector and the bias voltage response of the readout circuit are the most significant sources of non-uniformity.

[0003] In related technologies, non-uniformity correction can be performed by testing a reference source with uniform radiation in advance and collecting the corresponding puncture points in images under different test environments to achieve correction. The correction methods used can include single-point correction, two-point correction or multi-point correction. However, the above methods ignore the response time difference and can only be statically calibrated, which limits the dynamic application scenarios of infrared imaging and makes it difficult to apply to the mass production calibration of large-area array detectors. Summary of the Invention

[0004] This invention provides a non-uniformity correction method for point-by-point time-domain calibration of infrared detectors, which can achieve pixel-level response consistency reshaping under all operating conditions, providing underlying support for high frame rate and wide temperature range infrared imaging, and is suitable for mass production calibration of large-area array detectors.

[0005] The technical solution of the present invention is as follows: The non-uniformity correction method for point-by-point time-domain calibration of an infrared detector provided by the embodiments of the present invention includes: using a preset pulsed blackbody source as standard radiation, irradiating each pixel in the infrared detector point by point to obtain the simulated response data of each pixel in the infrared detector; constructing a simulated thermal response transfer function corresponding to each pixel based on the response data of each pixel in the infrared detector; and performing non-uniformity correction of the infrared detector based on the simulated thermal response transfer function corresponding to each pixel.

[0006] According to the above-described scheme, the beneficial effect of this invention is that, in the non-uniformity correction method for point-by-point time-domain calibration of infrared detectors provided by the embodiments of this invention, by using a preset pulsed blackbody source as standard radiation to illuminate each pixel in the infrared detector point by point, simulated response data of each pixel can be obtained. Based on the simulated response data of each pixel, a simulated thermal response transfer function corresponding to each pixel can be constructed. The simulated thermal response transfer function of each pixel can reflect the corresponding dynamic response mechanism. The simulated thermal response transfer function is corrected through the dynamic response mechanism, satisfying the dynamic application scenarios of infrared imaging, realizing pixel-level response consistency reshaping under all operating conditions, providing underlying support for high frame rate and wide temperature range infrared imaging, and is suitable for mass production calibration of large-area array detectors.

[0007] In some embodiments, the pulse width of the preset pulsed blackbody source is tp, and the time constant of the infrared detector is t0, 0.05*t0. <tp<0.1*t0。

[0008] In some embodiments, the step of using a preset pulsed blackbody source as standard radiation to illuminate each pixel in the infrared detector point by point to obtain simulated response data for each pixel in the infrared detector includes: using a preset pulsed blackbody source as standard radiation to illuminate each pixel in the infrared detector point by point at different temperatures to obtain simulated response data for each pixel in the infrared detector at different temperatures. The step of constructing a simulated thermal response transfer function corresponding to each pixel based on the response data of each pixel in the infrared detector includes: constructing a simulated thermal response transfer function corresponding to each pixel at different temperatures based on the simulated response data of each pixel in the infrared detector at different temperatures. The step of performing non-uniform correction on the infrared detector based on the simulated thermal response transfer function corresponding to each pixel includes: performing non-uniform correction on the infrared detector based on the ambient temperature and the simulated thermal response transfer function corresponding to each pixel at different temperatures.

[0009] In some embodiments, the step of using a preset pulsed blackbody source as standard radiation to illuminate each pixel in the infrared detector point by point at different temperatures to obtain simulated response data of each pixel in the infrared detector at different temperatures includes: illuminating each pixel in the infrared detector at a first frequency when the temperature is lower than or equal to a first preset temperature value; and illuminating each pixel in the infrared detector at a second frequency when the temperature is higher than the first preset temperature value, wherein the first frequency is higher than the second frequency.

[0010] In some embodiments, the step of using a preset pulsed blackbody source as standard radiation to illuminate each pixel in the infrared detector point by point at different temperatures to obtain simulated response data of each pixel in the infrared detector at different temperatures includes: sampling the simulated response data of each pixel in the infrared detector at a third frequency when the temperature is lower than or equal to a second preset temperature value; and sampling the simulated response data of each pixel in the infrared detector at a fourth frequency when the temperature is higher than the second preset temperature value, wherein the third frequency is lower than the fourth frequency.

[0011] In some implementations, the step of constructing the simulated thermal response transfer function for each pixel at different temperatures based on the simulated response data of each pixel in the infrared detector at different temperatures includes: constructing an original model of a first-order inertial element and a delayed response element; and optimizing the original model by fitting it with the least squares method based on the simulated response data of each pixel in the infrared detector at different temperatures to construct the simulated thermal response transfer function for each pixel at different temperatures.

[0012] In some implementations, the step of optimizing the undetermined parameters of the original model by fitting the simulated response data of each pixel in the infrared detector at different temperatures using the least squares method to construct the simulated thermal response transfer function corresponding to each pixel at different temperatures includes: during the iterative optimization of the model, using the Levenburg-Marquardt algorithm to minimize the sum of squared differences between the model predictions and the actual observations.

[0013] In some embodiments, the step of performing infrared detection non-uniformity correction based on the simulated thermal response transfer function corresponding to each pixel at different temperatures includes: determining the gain matrix and bias matrix corresponding to the infrared detector based on the simulated thermal response transfer function corresponding to each pixel at different temperatures, wherein the gain matrix is ​​used to correct the temporal responsivity difference of each pixel in the infrared detector, and the bias matrix is ​​used to correct the dark current and temperature drift difference of each pixel in the infrared detector.

[0014] In some embodiments, determining the gain matrix and bias matrix of the infrared detector based on the analog thermal response transfer function corresponding to each pixel at different temperatures includes: determining the gain matrix of the infrared detector based on the detection temperature of the infrared detector and the gain coefficient corresponding to the analog thermal response transfer function corresponding to each pixel at different temperatures; determining the dark current compensation term in the bias matrix based on the response voltage of each pixel in the infrared detector under no-irradiation conditions; determining the temperature drift compensation term in the bias matrix based on the temperature response gradient of each pixel in the infrared detector; and determining the bias matrix based on the dark current compensation term and the temperature drift compensation term in the bias matrix.

[0015] The present invention provides a non-uniform device for point-by-point temporal calibration of an infrared detector. The non-uniformity correction device includes a first processing module, a second processing module, and a third processing module. The first processing module is configured to use a preset pulsed blackbody source as standard radiation to illuminate each pixel in the infrared detector point-by-point, thereby obtaining analog response data for each pixel. The second processing module is configured to construct an analog thermal response transfer function corresponding to each pixel based on the response data of each pixel in the infrared detector. The third processing module is configured to perform non-uniformity correction of the infrared detector based on the analog thermal response transfer function corresponding to each pixel.

[0016] In the non-uniformity correction method for point-by-point time-domain calibration of infrared detectors provided in this invention, a preset pulsed blackbody source is used as standard radiation to illuminate each pixel in the infrared detector point by point, thereby obtaining the simulated response data of each pixel. Based on the simulated response data of each pixel, a simulated thermal response transfer function corresponding to each pixel can be constructed. The simulated thermal response transfer function of each pixel reflects the corresponding dynamic response mechanism. The simulated thermal response transfer function is corrected through the dynamic response mechanism, satisfying the dynamic application scenarios of infrared imaging, achieving pixel-level response consistency reshaping under all operating conditions, providing underlying support for high frame rate and wide temperature range infrared imaging, and is suitable for mass production calibration of large-area array detectors.

[0017] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0018] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which:

[0019] Figure 1 This is a flowchart illustrating a non-uniform correction method for point-by-point time-domain calibration of an infrared detector provided in certain embodiments of the present invention.

[0020] Figure 2 This is a flowchart illustrating a non-uniform correction method for point-by-point time-domain calibration of an infrared detector provided in certain embodiments of the present invention. Detailed Implementation

[0021] The present invention will now be further described with reference to the accompanying drawings and embodiments:

[0022] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are optional and are only used to explain the embodiments of the present invention, and should not be construed as limiting the embodiments of the present invention.

[0023] The non-uniformity of the spatial distribution of infrared images can severely affect the performance of infrared imaging systems and adversely impact their use. Therefore, non-uniformity correction is necessary to meet practical application requirements. The non-uniformity of infrared imaging systems mainly stems from the non-uniformity of the infrared detector response and the non-uniformity of the readout circuit electronics. Among these, the temperature response of the infrared detector and the bias voltage response of the readout circuit are the most significant sources of non-uniformity.

[0024] In related technologies, non-uniformity correction can be performed by testing a reference source with uniform radiation in advance and collecting the corresponding puncture points in images under different test environments to achieve correction. The correction methods used can include single-point correction, two-point correction or multi-point correction. However, the above methods ignore the response time difference and can only be statically calibrated, which limits the dynamic application scenarios of infrared imaging and makes it difficult to apply to the mass production calibration of large-area array detectors.

[0025] To address the aforementioned technical problems, embodiments of the present invention provide a method and apparatus for non-uniformity correction of point-by-point time-domain calibration of infrared detectors.

[0026] Reference Figure 1 The non-uniformity correction method for point-by-point time-domain calibration of an infrared detector provided by the embodiments of the present invention includes:

[0027] Step 01: Using a preset pulsed blackbody source as standard radiation, illuminate each pixel in the infrared detector point by point to obtain the simulated response data of each pixel in the infrared detector.

[0028] Step 02: Based on the response data of each pixel in the infrared detector, construct the simulated thermal response transfer function corresponding to each pixel;

[0029] Step 03: Perform non-uniform correction of the infrared detector based on the simulated thermal response transfer function corresponding to each pixel.

[0030] The non-uniformity correction method for point-by-point time-domain calibration of infrared detectors provided by the embodiments of the present invention uses a preset pulsed blackbody source as standard radiation to illuminate each pixel in the infrared detector point by point, thereby obtaining the simulated response data of each pixel. Based on the simulated response data of each pixel, the simulated thermal response transfer function corresponding to each pixel can be constructed.

[0031] The thermal relaxation process of an infrared detector mainly refers to the process by which its sensitive element absorbs infrared radiation, causing changes in physical parameters (such as resistance and voltage) due to temperature variations. The simulated response data for each pixel includes the response data corresponding to the thermal relaxation process of the infrared detector, ensuring that the simulated thermal response transfer function for each pixel can fully reflect the process of temperature-induced changes in physical parameters (such as resistance and voltage). During the actual operation of the infrared detector, non-uniform correction is performed on the pixels to be processed based on the simulated thermal response transfer function.

[0032] The simulated thermal response transfer function of each pixel can reflect the corresponding dynamic response mechanism. The simulated thermal response transfer function is corrected through the dynamic response mechanism, which meets the dynamic application scenarios of infrared imaging, realizes the pixel-level response consistency reshaping under all working conditions, provides underlying support for high frame rate and wide temperature range infrared imaging, and is suitable for mass production calibration of large array detectors.

[0033] In some implementations, the pulse width of the preset pulsed blackbody source is tp, and the time constant of the infrared detector is t0, 0.05*t0. <tp<0.1*t0。

[0034] Specifically, the pulse width setting of the pulsed blackbody source must be strictly matched with the thermodynamic characteristics and engineering constraints of the infrared detector to meet the actual application scenarios.

[0035] The time constant of an infrared detector is t0, defined as the time required for the detector's temperature response to reach 63.2% of its steady-state change. It is determined by both heat capacity (Cth) and thermal conductivity (Gth): t0 = Gth / Cth. Generally, t0 can be taken as 5ms-20ms.

[0036] For different time constants t0, the pulse width tp of the preset pulsed blackbody source is set differently. For low thermal conductivity designs (such as vacuum packaging), the time constant t0 is longer (15-20ms), so the pulse width tp of the preset pulsed blackbody source is set wider to cover the thermal relaxation process. For high thermal conductivity designs (such as non-vacuum packaging), the time constant t0 is shorter (5-10ms), so the pulse width tp of the preset pulsed blackbody source is set narrower to avoid energy accumulation.

[0037] The pulse width tp needs to be greater than 0.05*t0 to ensure that it covers the initial response stage of the thermal relaxation process and that the pulse width tp can capture the rising edge of the response curve.

[0038] The pulse width tp needs to be less than 0.1*t0 to avoid the infrared detector temperature approaching steady state, which could lead to excessively high microbridge temperature and cause self-immolation.

[0039] Reference Figure 2In some implementations, step 01: using a preset pulsed blackbody source as standard radiation, each pixel in the infrared detector is illuminated point by point to obtain analog response data of each pixel in the infrared detector, including:

[0040] Step 011: At different temperatures, a preset pulsed blackbody source is used as the standard radiation to illuminate each pixel in the infrared detector point by point, so as to obtain the simulated response data of each pixel in the infrared detector at different temperatures.

[0041] Step 02: Based on the response data of each pixel in the infrared detector, construct the simulated thermal response transfer function corresponding to each pixel, including:

[0042] Step 021: Based on the simulated response data of each pixel in the infrared detector at different temperatures, construct the simulated thermal response transfer function corresponding to each pixel at different temperatures.

[0043] Step 03: Perform non-uniformity correction on the infrared detector based on the analog thermal response transfer function corresponding to each pixel, including:

[0044] Step 031: Perform non-uniform correction of the infrared detector based on the ambient temperature and the simulated thermal response transfer function corresponding to each pixel at different temperatures.

[0045] Specifically, the embodiments of this application generate non-uniform correction parameters at different temperature points through multi-dimensional parameter mapping, supporting dynamic non-uniform correction of infrared detectors based on different temperatures.

[0046] In step 011, a preset pulsed blackbody source can be used as standard radiation at different temperatures to obtain simulated response data. For example, within a temperature range of -40℃ to 85℃, a temperature test point can be calibrated every 5℃, for a total of 25 temperature test points. At each of the different temperature test points, the preset pulsed blackbody source is used as standard radiation to illuminate each pixel in the infrared detector point by point, thereby obtaining 25 sets of simulated response data A1 to A25 for each pixel in the infrared detector at the 25 temperature test points.

[0047] The simulated response data can be denoted as [T, Vbias, position (i, j), response, timestamp]. The position (i, j) can be denoted as the pixel in the i-th row and j-th column, T is the temperature, and Vbias is the bias voltage.

[0048] In step 021, the response curves are fitted according to the 25 sets of simulated response data A1 to A25 corresponding to the 25 temperature test points to obtain the simulated thermal response transfer function corresponding to each pixel under the 25 temperature test points. The 25 sets of simulated thermal response transfer functions are denoted as B1 to B25.

[0049] In step 022, based on the 25 sets of simulated thermal response transfer functions denoted as B1 to B25, corresponding 25 sets of correction parameters C1 to C25 can be set. During actual operation, the correction circuit determines the matching correction parameters based on the actual ambient temperature, and then performs non-uniform correction on each pixel according to the matching correction parameters.

[0050] For example, a Gain / Offset register can be integrated into the ROIC to perform non-uniform compensation in real time according to a formula:

[0051] Voutcorr(i,j) = Kij * Vraw(i,j) + Offsetij, where Voutcorr(i,j) is the response voltage after non-uniform compensation for the pixel in row i and column j, Kij is the gain coefficient for the pixel in row i and column j, Vraw(i,j) is the original response voltage for the pixel in row i and column j, and Offsetij is the compensation voltage for the pixel in row i and column j.

[0052] If the temperature difference between the actual ambient temperature and the temperature test point is less than 2℃, the calibration parameter corresponding to the temperature test point can be determined as the matching calibration parameter. For example, if the temperature test point corresponding to calibration parameter C25 is 85℃, and the ambient temperature is 84℃, calibration parameter C25 can be determined as the matching calibration parameter, and real-time non-uniform compensation can be performed based on the above formula according to calibration parameter C25.

[0053] In some embodiments, real-time non-uniform compensation can also be performed based on multiple matching correction parameters. For example, if the temperature test point corresponding to correction parameter C25 is 85℃ and the temperature test point corresponding to correction parameter C24 is 80℃, and the ambient temperature is 82℃, correction parameters C24 and C25 can be determined as matching correction parameters, and a new target correction parameter Cout = K1*C24 + K2*C25 can be determined based on correction parameters C24 and C25, where K1 can be 0.6 and K2 can be 0.4. Finally, real-time non-uniform compensation is performed based on the above formula according to the target correction parameter Cout.

[0054] In some implementations, step 011: At different temperatures, a preset pulsed blackbody source is used as standard radiation to illuminate each pixel in the infrared detector point by point, in order to obtain simulated response data of each pixel in the infrared detector at different temperatures, including:

[0055] When the temperature is lower than or equal to the first preset temperature value, each pixel in the infrared detector is irradiated at the first frequency.

[0056] When the temperature is higher than the first preset temperature value, each pixel in the infrared detector is illuminated at a second frequency, where the first frequency is higher than the second frequency.

[0057] Specifically, when the temperature is higher than the first preset temperature value, the thermal diffusivity of the pixel increases significantly (thermal conductivity), the heat transfer between adjacent pixels is accelerated, and the output response of adjacent units is coupled and interfered. The temperature rise of the activated pixel will cause the temperature of the neighboring pixel to rise, resulting in signal response distortion. At this time, the frequency of each pixel in the infrared detector can be reduced from the first frequency to the second frequency, and the interval time of each pixel can be increased to avoid the thermal coupling superposition effect.

[0058] In some embodiments, the first preset temperature value can be set to 60°C, and each pixel in the infrared detector is irradiated at a second frequency to ensure that the time interval between each pixel in the infrared detector is greater than 5 times the pixel time constant.

[0059] In some implementations, step 011: At different temperatures, a preset pulsed blackbody source is used as standard radiation to illuminate each pixel in the infrared detector point by point, in order to obtain simulated response data of each pixel in the infrared detector at different temperatures, including:

[0060] When the temperature is lower than or equal to the second preset temperature value, the analog response data of each pixel in the infrared detector is sampled at a third frequency.

[0061] When the temperature is higher than the second preset temperature value, the analog response data of each pixel in the infrared detector is sampled according to the fourth frequency, and the third frequency is lower than the fourth frequency.

[0062] Specifically, when the temperature is below the second preset temperature value, the pixel carrier mobility decreases, and its power spectral density is inversely proportional to the frequency. The low-frequency noise amplitude increases significantly, manifesting as slow-varying drift interference in the output signal. At this time, the sampling frequency of the analog response data of each pixel in the sampling infrared detector can be reduced from the third frequency to the fourth frequency, increasing the sampling interval. Differential cancellation can be achieved by utilizing the uncorrelated characteristics of 1 / f noise, thereby improving the noise suppression rate in the low-temperature region.

[0063] In some embodiments, the second preset temperature value can be set to 0°C.

[0064] In some implementations, step 021: Based on the simulated response data of each pixel in the infrared detector at different temperatures, construct the simulated thermal response transfer function corresponding to each pixel at different temperatures, including:

[0065] Construct the original model of the first-order inertial element and the delayed response element;

[0066] Based on the simulated response data of each pixel in the infrared detector at different temperatures, the original model is optimized by fitting the least squares method to construct the simulated thermal response transfer function of each pixel at different temperatures.

[0067] Specifically, the thermal response transfer function of a pixel is a primitive model of a first-order inertial element and a delayed response element, and its mathematical expression is as follows:

[0068]

[0069] Where K is the gain coefficient, T is the thermal time constant, and τ is the transmission delay, all three collectively characterizing the pixel's response properties. Based on the original model described above, the time-domain response equation is obtained through the inverse Laplace transform, and an experimental model incorporating delay and noise is constructed:

[0070]

[0071] Where τd represents the circuit delay (in μs), and n(t) represents the superimposed noise (including Johnson noise, 1 / f noise, etc.).

[0072] In some implementations, based on the simulated response data of each pixel in the infrared detector at different temperatures, the undetermined parameters of the original model are optimized by fitting using the least squares method to construct the simulated thermal response transfer function corresponding to each pixel at different temperatures, including:

[0073] During the iterative optimization of the model, the Levenburg-Marquardt algorithm is used to minimize the sum of squared differences between the model's predicted values ​​and the actual observed values.

[0074] Specifically, the Levenburg-Marquardt algorithm can be used to solve for (K, T, τ) to obtain the pixel-specific gain-bias correction coefficient matrix.

[0075] The Levenberg-Marquardt algorithm is used to minimize the sum of squared residuals:

[0076]

[0077] Initialization parameter settings:

[0078]

[0079] Vsteady is the steady-state voltage of the simulated response curve, and trise is the rise time of the simulated response curve.

[0080] Calculate the Jacobian matrix J, with the following elements:

[0081]

[0082] Update parameter: Δβ=(J T J+λI) -1 J T r

[0083] Where r is the residual vector.

[0084] Convergence condition:

[0085] ||Δβ||<10 -6

[0086] In some implementations, step 031: performing infrared detection non-uniformity correction based on the simulated thermal response transfer function corresponding to each pixel at different temperatures, including:

[0087] Based on the simulated thermal response transfer function corresponding to each pixel at different temperatures, the gain matrix and bias matrix of the infrared detector are determined. The gain matrix is ​​used to correct the time domain responsivity differences of each pixel in the infrared detector, and the bias matrix is ​​used to correct the dark current and temperature drift differences of each pixel in the infrared detector.

[0088] Specifically, the corrected pixel output is generated by a combination of gain compensation and offset compensation:

[0089] V corr (i,j)=G ij ·V raw (i,j)+O ij

[0090] Gij can be the gain coefficient corresponding to the pixel in the i-th row and j-th column, and Oij can be the bias coefficient corresponding to the pixel in the i-th row and j-th column.

[0091] Based on the solution of (K, T, τ), the gain matrix and bias matrix corresponding to the infrared detector can be generated. The gain matrix corresponding to the infrared detector can include the gain coefficient corresponding to each pixel, and the bias matrix corresponding to the infrared detector can include the bias coefficient corresponding to each pixel.

[0092] In some implementations, the gain matrix and bias matrix of the infrared detector are determined based on the simulated thermal response transfer function corresponding to each pixel at different temperatures, including:

[0093] The gain matrix of the infrared detector is determined based on the detection temperature of the infrared detector and the gain coefficient of the analog thermal response transfer function corresponding to each pixel at different temperatures.

[0094] Based on the response voltage of each pixel in the infrared detector under no-irradiation conditions, the dark current compensation term in the bias matrix is ​​determined.

[0095] The temperature drift compensation term in the bias matrix is ​​determined based on the temperature response gradient of each pixel in the infrared detector.

[0096] The bias matrix is ​​determined based on the dark current compensation term and the temperature drift compensation term in the bias matrix.

[0097] Specifically, within a set of test temperatures, the average gain of the entire array can be determined based on the simulated thermal response transfer function corresponding to each pixel. Using the average gain of the entire array as a benchmark, the gain coefficient of each pixel can be calculated.

[0098] Average gain of the entire array:

[0099]

[0100] Where M represents that the pixel array has M rows, N represents that the pixel array has N columns, and Kij represents the solution of the gain coefficient K in the simulated thermal response transfer function corresponding to the pixel in the i-th row and j-th column.

[0101] Gain coefficient per pixel:

[0102]

[0103] In some embodiments, in order to eliminate response rate differences, the obtained Gij can be limited (e.g., 0.5 ≤ Gij ≤ 2.0) to avoid distortion due to extreme values.

[0104] The bias coefficient includes a dark current compensation term. Under no-irradiation conditions, the pixel dark output Vdark(i,j) is measured, and the dark current compensation term is:

[0105] O ij =-G ij ·V dark (i,j)

[0106] Dark current compensation is used to eliminate fixed bias noise.

[0107] The temperature drift compensation term can be determined based on the analog thermal response transfer function corresponding to each pixel:

[0108]

[0109] Where α is the temperature drift coefficient, Tij represents the solution of the temperature constant T in the simulated thermal response transfer function corresponding to the i-th row and j-th column pixel, and Tenv can be the measured ambient temperature.

[0110] Based on the above formula, the gain coefficient and bias coefficient of each pixel at different temperatures can be determined, and then the gain matrix and bias matrix of the infrared detector at different temperatures can be determined. The size of the generated matrix is ​​the same as that of the infrared detector array (e.g., 640×512).

[0111] In some embodiments, during the process of illuminating each pixel in the infrared detector point by point to obtain the analog response data of each pixel in the infrared detector, a specific sampling technique may be used to suppress noise during the signal acquisition stage.

[0112] For example, oversampling and temporal truncation techniques can be used. By oversampling at a high multiple (e.g., sampling frequency ≥ 32 times the pixel output frequency), a large number of samples are generated. Then, distorted samples are removed. A distorted sample template is established through multi-frame statistics to remove transient interference. Then, statistical distribution analysis is performed on the remaining samples to remove outlier samples that deviate from the mean ± 3σ, ensuring that the number of samples is a power of 2 (e.g., 32 / 64). Mean filtering or median filtering is then used to avoid quantization errors.

[0113] For example, Correlated Double Sampling (CDS) technology can be used to sample each pixel twice within the pulsed blackbody illumination period. The first sampling is a reset level sampling, which is used to capture reset noise (KTC noise, 1 / f noise). The second sampling is a signal level sampling, which includes the superposition value of signal and noise. By differential processing of the two sampling results, time-dependent noise (such as KTC noise and fixed pattern noise) can be eliminated.

[0114] In some embodiments, an improved Non-Local Means (NLM) algorithm can be used to denoise the sampled data. The improvement of the NLM algorithm mainly focuses on two aspects: optimizing computational efficiency and enhancing structure fidelity. It uniformly compresses the three channels (θ, φ, g) of the input features to half the original number of channels, significantly reducing the computational cost of convolution. Max pooling layers (such as 2×2 pooling) are added to the φ and g paths, reducing the computational cost of pixel neighborhood comparison to 1 / 4 of the original. Mathematically, this can be expressed as:

[0115]

[0116] Here, θi and φj are the pooled feature blocks, and only 1 / 4 of the pixel pairs need to be calculated after sampling.

[0117] For each pixel position i, calculate the SSIM structural similarity. The final weight is the product of grayscale similarity and structural similarity, as shown in the formula:

[0118]

[0119] When calculating SSIM, the entire image needs to be processed block by block. p and q represent two sets of response data extracted at the same pixel unit location. μp and μq can represent the pixel mean (luminance information) of regions p and q, respectively. σp and σq can represent the standard deviation (contrast information) of the two regions, respectively, and σpq represents the covariance (structural similarity, quantifying the correlation between pixel changes in the two regions).

[0120] The similarity is defined by the squared difference in gray levels of neighboring blocks (Euclidean distance):

[0121]

[0122] Use an exponential function to convert distance into pixel weights (Wpixel):

[0123]

[0124] Where h is the filter strength parameter, which controls the weight decay rate.

[0125] Wtotal (total weight) is multiplied by the pixel weight (Wpixel) by introducing structural similarity weight (SSIM):

[0126] W total =W pixel ×SSIM(p,q)

[0127] The corrected output value for each target pixel i is calculated by the weighted average of all pixels in its neighborhood:

[0128]

[0129] The noise of all pixel units can be corrected using Wtotal (total weight).

[0130] The non-uniformity correction device provided in the embodiments of the present invention includes a memory and a processor. The memory is configured to store a computer program, and when the processor executes the computer program, it implements the non-uniformity correction method provided in the embodiments of the present invention.

[0131] The non-uniformity correction method provided in this embodiment of the invention can be implemented by a non-uniformity correction device; that is, the non-uniformity correction device is used to implement the non-uniformity correction method. Of course, in other embodiments, the non-uniformity correction method can also be implemented by other devices or equipment, and is not limited to being implemented by a non-uniformity correction device. The non-uniformity correction device may not be exclusively used to implement the non-uniformity correction method of this invention, but can implement other functions or methods.

[0132] In some embodiments, the non-uniformity correction device includes a first processing module, a second processing module, and a third processing module. The first processing module is configured to use a preset pulsed blackbody source as standard radiation to illuminate each pixel in the infrared detector point by point, thereby acquiring analog response data for each pixel in the infrared detector. The second processing module is configured to construct an analog thermal response transfer function corresponding to each pixel based on the response data of each pixel in the infrared detector. The third processing module is configured to perform non-uniformity correction on the infrared detector based on the analog thermal response transfer function corresponding to each pixel.

[0133] The non-uniform correction device provided in this application includes all the technical effects of the non-uniform correction method provided in this application, which will not be elaborated here.

[0134] In the description of this specification, the references to terms such as "some embodiments," "in one example," and "exemplarily" indicate that a specific feature, structure, material, or characteristic described in connection with an embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0135] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process, and the scope of the preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of the invention pertain.

[0136] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are optional and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A non-uniform correction method for point-by-point time-domain calibration of an infrared detector, characterized in that, The method includes: A preset pulsed blackbody source is used as the standard radiation to illuminate each pixel in the infrared detector point by point in order to obtain the simulated response data of each pixel in the infrared detector. Based on the response data of each pixel in the infrared detector, a simulated thermal response transfer function corresponding to each pixel is constructed; The non-uniformity correction of the infrared detector is performed based on the simulated thermal response transfer function corresponding to each pixel.

2. The non-uniform correction method for point-by-point time-domain calibration of an infrared detector according to claim 1, characterized in that, The pulse width of the preset pulsed blackbody source is tp, and the time constant of the infrared detector is t0, 0.05*t0. <tp<0.1*t0。 3. The non-uniform correction method for point-by-point time-domain calibration of an infrared detector according to claim 1 or 2, characterized in that, The process of using a preset pulsed blackbody source as standard radiation to illuminate each pixel in the infrared detector point by point to obtain simulated response data for each pixel in the infrared detector includes: At different temperatures, a preset pulsed blackbody source is used as the standard radiation to illuminate each pixel in the infrared detector point by point in order to obtain the simulated response data of each pixel in the infrared detector at different temperatures. The step of constructing a simulated thermal response transfer function for each pixel based on the response data of each pixel in the infrared detector includes: Based on the simulated response data of each pixel in the infrared detector at different temperatures, a simulated thermal response transfer function corresponding to each pixel at different temperatures is constructed. The step of performing non-uniform correction of the infrared detector based on the analog thermal response transfer function corresponding to each pixel includes: The non-uniformity correction of the infrared detector is performed based on the ambient temperature and the simulated thermal response transfer function corresponding to each pixel at different temperatures.

4. The non-uniform correction method for point-by-point time-domain calibration of an infrared detector according to claim 3, characterized in that, The process involves using a preset pulsed blackbody source as standard radiation at different temperatures to illuminate each pixel in the infrared detector point by point, in order to obtain simulated response data of each pixel in the infrared detector at different temperatures, including: When the temperature is lower than or equal to a first preset temperature value, each pixel in the infrared detector is illuminated at a first frequency. When the temperature is higher than the first preset temperature value, each pixel in the infrared detector is illuminated at a second frequency, where the first frequency is higher than the second frequency.

5. The non-uniform correction method for point-by-point time-domain calibration of an infrared detector according to claim 3, characterized in that, The process involves using a preset pulsed blackbody source as standard radiation at different temperatures to illuminate each pixel in the infrared detector point by point, in order to obtain simulated response data of each pixel in the infrared detector at different temperatures, including: When the temperature is lower than or equal to a second preset temperature value, the analog response data of each pixel in the infrared detector is sampled at a third frequency. When the temperature is higher than the second preset temperature value, the analog response data of each pixel in the infrared detector is sampled at a fourth frequency, wherein the third frequency is lower than the fourth frequency.

6. The non-uniform correction method for point-by-point time-domain calibration of an infrared detector according to claim 3, characterized in that, The step of constructing a simulated thermal response transfer function for each pixel at different temperatures based on the simulated response data of each pixel in the infrared detector includes: Construct the original model of the first-order inertial element and the delayed response element; Based on the simulated response data of each pixel in the infrared detector at different temperatures, the original model is optimized by fitting using the least squares method to construct the simulated thermal response transfer function corresponding to each pixel at different temperatures.

7. The non-uniform correction method for point-by-point time-domain calibration of an infrared detector according to claim 6, characterized in that, The step of constructing a simulated thermal response transfer function for each pixel at different temperatures by fitting and optimizing the undetermined parameters of the original model using the least squares method based on the simulated response data of each pixel in the infrared detector at different temperatures includes: During the iterative optimization of the model, the Levenburg-Marquardt algorithm is used to minimize the sum of squared differences between the model's predicted values ​​and the actual observed values.

8. The non-uniform correction method for point-by-point time-domain calibration of an infrared detector according to claim 3, characterized in that, The step of performing infrared detection non-uniformity correction based on the simulated thermal response transfer function corresponding to each pixel at different temperatures includes: Based on the simulated thermal response transfer function corresponding to each pixel at different temperatures, the gain matrix and bias matrix corresponding to the infrared detector are determined. The gain matrix is ​​used to correct the time domain responsivity differences of each pixel in the infrared detector, and the bias matrix is ​​used to correct the dark current and temperature drift differences of each pixel in the infrared detector.

9. The non-uniform correction method for point-by-point time-domain calibration of an infrared detector according to claim 8, characterized in that, The step of determining the gain matrix and bias matrix of the infrared detector based on the simulated thermal response transfer function corresponding to each pixel at different temperatures includes: The gain matrix corresponding to the infrared detector is determined based on the detection temperature of the infrared detector and the gain coefficient corresponding to the analog thermal response transfer function of each pixel at different temperatures. The dark current compensation term in the bias matrix is ​​determined based on the response voltage of each pixel in the infrared detector under no-irradiation conditions. The temperature drift compensation term in the bias matrix is ​​determined based on the temperature response gradient of each pixel in the infrared detector. The bias matrix is ​​determined based on the dark current compensation term and the temperature drift compensation term in the bias matrix.

10. A non-uniform device for point-by-point time-domain calibration of an infrared detector, characterized in that, The non-uniformity correction device includes: The first processing module is configured to use a preset pulsed blackbody source as standard radiation to illuminate each pixel in the infrared detector point by point in order to obtain the analog response data of each pixel in the infrared detector. The second processing module is configured to construct an analog thermal response transfer function for each pixel based on the response data of each pixel in the infrared detector. The third processing module is configured to perform non-uniform correction of the infrared detector based on the analog thermal response transfer function corresponding to each pixel.

Images