Laser wavefront distortion correction system and method based on adaptive optics

Abstract

The invention relates to the technical field of image processing, and discloses a laser wavefront distortion correction system and method based on adaptive optics, and the system comprises a first module which is used for calculating a noise covariance matrix and an interaction matrix; the second module is used for solving a control matrix and a phase reconstruction matrix; the third module is used for calculating to obtain a current slope vector; the fourth module is used for updating a deformable mirror command; the fifth module is used for calculating to obtain a point spread function; and the sixth module is used for calculating an image quality index and outputting a correction result. According to the method, a noise covariance matrix and an interaction matrix are accurately constructed under reference light, and a weighted least square method with double regularization is combined, so that the influence of measurement noise is effectively inhibited, the stroke constraint of an actuator is balanced, and the robustness of a correction model is greatly improved; by means of combination of integral control, leakage processing and amplitude limiting design, saturation of an actuator due to command accumulation is avoided, and correction error superposition is prevented.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and more specifically, to a laser wavefront distortion correction system and method based on adaptive optics. Background Technology

[0002] Lasers, with their advantages of high directionality and high monochromaticity, are widely used in imaging, ranging, target recognition and other fields. These scenarios place stringent requirements on the clarity and stability of laser imaging quality. During laser transmission, environmental factors such as atmospheric turbulence and temperature fluctuations can disrupt the consistency of the wavefront phase. At the same time, assembly and adjustment errors and component defects in the optical system can also lead to wavefront distortion, causing the beam energy distribution to shift and the sub-aperture centroid position to change irregularly, thereby introducing measurement noise.

[0003] In existing wavefront distortion correction techniques, some methods do not fully consider the correlation of noise, and only use a simple diagonal weighted model, resulting in noise significantly interfering with the correction results. The correction matrix construction lacks effective regularization processing, making it difficult to balance measurement error and actuator travel constraints, and easily leading to high-gain oscillations or insufficient correction. The phase reconstruction process does not incorporate smoothing priors, and the reconstruction results are greatly affected by noise, failing to accurately reflect the true residual phase. In addition, existing technologies mostly rely on a single index to judge the correction effect, lacking intuitive and unified quantitative standards, making it difficult to reproduce and accept the correction results. At the same time, the actuator is prone to saturation due to command accumulation during the correction process, further reducing the correction stability and severely restricting the application of lasers in high-precision scenarios. Summary of the Invention

[0004] This invention provides a laser wavefront distortion correction system and method based on adaptive optics, which solves the technical problems mentioned in the background.

[0005] This invention provides a laser wavefront distortion correction system based on adaptive optics, comprising: The first module is used to obtain the reference centroid coordinates and current centroid coordinates of each sub-aperture through a wavefront sensor under reference light illumination, convert the slope to form a slope vector, calculate the noise covariance matrix, apply small displacement commands to each actuator of the deformable mirror, and record the slope increment to form an interaction matrix. The second module is used to solve the control matrix and phase reconstruction matrix using regularized weighted least squares based on the interaction matrix and the noise covariance matrix. The third module is used to obtain the current centroid coordinates and calculate the slope after preprocessing the wavefront sensor image under working light illumination, and to obtain the current slope vector by subtracting the static bias. The fourth module is used to obtain the correction amount by mapping the current slope vector according to the control matrix, update the deformable mirror command through integral control and leakage, and satisfy the actuator stroke constraint through amplitude limiting projection. The fifth module is used to calculate the residual slope between the current slope vector and the predicted slope, which is then converted into the residual phase by the phase reconstruction matrix. Combined with the aperture function, the point spread function is obtained to evaluate the imaging quality. The sixth module is used to calculate the image quality index based on the residual phase. When the preset threshold is reached or the residual meets the requirements, the deformable mirror command is frozen and the correction result is output.

[0006] Further, under reference light illumination, for each sub-aperture, the reference pixel intensity of each pixel within that sub-aperture is obtained. The x-coordinate of each pixel is multiplied by the corresponding reference pixel intensity, and the sum is obtained to get the first numerator. The sum of all reference pixel intensities within that sub-aperture is obtained to get the first denominator. The first numerator is divided by the first denominator to obtain the reference x-direction centroid coordinates of that sub-aperture. The y-coordinate of each pixel is multiplied by the corresponding reference pixel intensity, and the sum is obtained to get the second numerator. The second numerator is divided by the first denominator to obtain the reference y-direction centroid coordinates of that sub-aperture. The same method is used to calculate... Obtain the current centroid coordinates of each sub-aperture; subtract the reference x-axis centroid coordinate of each sub-aperture from its current x-axis centroid coordinate, and divide the difference by the focal length of the microlens to obtain the x-axis slope of that sub-aperture; subtract the reference y-axis centroid coordinate of each sub-aperture from its current y-axis centroid coordinate, and divide the difference by the focal length of the microlens to obtain the y-axis slope of that sub-aperture; arrange the x-axis slopes of all sub-apertures in a preset order, and then arrange the y-axis slopes of all sub-apertures in the same preset order to form a slope vector.

[0007] Furthermore, under the condition that the reference light remains static, slope vectors are acquired multiple times, and the average value of the slope at corresponding positions in all groups of slope vectors is calculated to obtain the mean value of the slope vectors. The mean value of the slope vector at each position in each group of slope vectors is subtracted from the mean value of the slope vector at the corresponding position to obtain the deviation vector corresponding to each group of slope vectors. Each group of deviation vectors is multiplied by its own transpose matrix to obtain the deviation product matrix for each group. The average value of all groups of deviation product matrices is calculated to obtain the noise covariance matrix. For each actuator of the deformable mirror, a preset micro-pressure is applied to the actuator. Small displacement command; record the slope vector before applying the small displacement command as the pre-application slope vector; record the slope vector after applying the small displacement command as the post-application slope vector; subtract the slope of the corresponding position in the pre-application slope vector from the slope of each position in the post-application slope vector to obtain the slope increment corresponding to the actuator; divide each position increment in the slope increment by the small displacement command of the actuator to obtain the matrix column corresponding to the actuator; arrange the matrix columns corresponding to all actuators in the preset numbering order of the actuators to form an interaction matrix.

[0008] Furthermore, the control matrix is ​​calculated using a weighted least squares method with regularization, including the following steps: Step S201: Calculate the product of the transpose of the interaction matrix and the inverse of the noise covariance matrix; Step S202: Multiply the product by the interaction matrix to obtain the first intermediate calculation result; Step S203: Calculate the product of the regularization coefficient of the control matrix and the identity matrix to obtain the second intermediate calculation result; Step S204: Add the first intermediate calculation result to the second intermediate calculation result to obtain the third intermediate calculation result; Step S205: Calculate the inverse result of the third intermediate calculation result, and multiply the inverse result with the transpose of the interaction matrix and the inverse of the noise covariance matrix in sequence to obtain the control matrix. The control matrix converts the slope vector into deformable mirror commands.

[0009] Furthermore, the phase reconstruction matrix is ​​calculated using a weighted least squares method with regularization, including the following steps: Step S301: Calculate the product of the transpose of the gradient operator and the inverse of the noise covariance matrix; Step S302: Multiply the product obtained in the first step with the gradient operator to obtain the fourth intermediate calculation result; Step S303: Calculate the product of the transpose of the smoothing prior operator and the smoothing prior operator; Step S304: Calculate the product of the regularization coefficient of the phase reconstruction matrix and the product obtained in the third step to obtain the fifth intermediate calculation result; Step S305: Add the fourth intermediate calculation result to the fifth intermediate calculation result to obtain the sixth intermediate calculation result; Step S306: Calculate the inverse of the sixth intermediate calculation result, and multiply the inverse result by the transpose of the gradient operator and the inverse of the noise covariance matrix in sequence to obtain the phase reconstruction matrix.

[0010] Furthermore, for each pixel of each sub-aperture, the difference between the working light pixel intensity of that pixel and the dark field value of the corresponding sub-aperture is calculated. It is determined whether the difference is greater than or equal to a preset threshold. If it is greater than or equal to the preset threshold, the threshold mask is set to one; if it is less than the preset threshold, the threshold mask is set to zero. The above difference is multiplied by the threshold mask to obtain the preprocessed pixel intensity of each pixel of each sub-aperture. For each sub-aperture, calculate the sum of the products of the lateral coordinates of all pixels in that sub-aperture and the corresponding preprocessed pixel intensities, calculate the sum of the intensities of all preprocessed pixels in that sub-aperture, divide the sum of the products by the sum of the intensities to obtain the current centroid lateral coordinates of that sub-aperture; using the same method, calculate the current centroid longitudinal coordinates of that sub-aperture by using the pixel longitudinal coordinates and the corresponding preprocessed pixel intensities. For each sub-aperture, calculate the difference between the current centroid lateral coordinate and the reference centroid lateral coordinate of the sub-aperture, divide the difference by the microlens focal length to obtain the lateral slope of the sub-aperture; calculate the difference between the current centroid longitudinal coordinate and the reference centroid longitudinal coordinate of the sub-aperture, divide the difference by the microlens focal length to obtain the longitudinal slope of the sub-aperture; concatenate all lateral and longitudinal slopes in the order of lateral slopes of all sub-apertures arranged sequentially, followed by longitudinal slopes of all sub-apertures arranged sequentially after the lateral slopes, to form a slope vector without offset subtraction; The current slope vector is obtained by subtracting the slope vector without bias subtraction from the slope vector with static bias.

[0011] Furthermore, the integral gain is multiplied by the control matrix, the resulting product is multiplied by the current slope vector, and the result is negative to obtain the correction vector. Subtract the leakage coefficient from 1, multiply the difference by the current deformable mirror command vector, and add the product to the correction vector to obtain the pre-updated deformable mirror command vector. For each component of the pre-updated deformable mirror command vector, compare the component with the corresponding component of the lower bound vector of the executor command and the corresponding component of the upper bound vector of the executor command. If the component is less than the corresponding component of the lower bound vector of the executor command, then take the corresponding component of the lower bound vector of the executor command as the component at that position. If the component is greater than the corresponding component of the upper bound vector of the executor command, then take the corresponding component of the upper bound vector of the executor command as the component at that position. If the component is between the corresponding components of the lower bound vector of the executor command and the corresponding components of the upper bound vector of the executor command, then keep the component unchanged to obtain the bounded deformable mirror command vector.

[0012] Further, the interaction matrix is ​​multiplied with the current deformable mirror command vector to obtain the predicted slope vector; the current slope vector is subtracted from the predicted slope vector to obtain the residual slope vector; the phase reconstruction matrix is ​​multiplied with the residual slope vector to obtain the residual phase; the imaginary unit is multiplied with the residual phase, the product is exponentially multiplied, and the result is multiplied with the aperture function to obtain the aperture complex amplitude; a Fourier transform is performed on the aperture complex amplitude, the modulus of the Fourier transform result is taken, and the result is squared to obtain the point spread function.

[0013] Furthermore, the residual phases corresponding to all sampling points in the aperture sampling point set are summed, and the sum is divided by the number of aperture sampling points to obtain the average value of the residual phases. For the residual phase of each sampling point in the aperture sampling point set, calculate the difference between the residual phase and the average residual phase, square each difference, sum all the squared results, divide the sum by the number of aperture sampling points, and take the square root of the quotient to obtain the root mean square error of the residual wavefront. Calculate the ratio of 2π to the working wavelength, square the ratio, calculate the square of the root mean square of the residual wavefront error, multiply the results of the two square operations, take the negative value of the product, and perform an exponential operation on the negative value to obtain the image quality index. Squaring each component of the residual slope vector, summing all the squaring results, and taking the square root of the sum to obtain the L2 norm of the residual slope vector. If the image quality index is greater than or equal to the image quality index threshold, or the L2 norm of the residual slope vector is less than or equal to the residual slope vector threshold, then the stopping condition is met. When the stopping condition is met, the current deformable mirror command vector is determined as the frozen deformable mirror command vector, and the frozen deformable mirror command vector and the corresponding correction result are output.

[0014] This invention provides a laser wavefront distortion correction method based on adaptive optics, comprising the following steps: Step S401: Under the illumination of the reference light, the reference centroid coordinates and the current centroid coordinates of each sub-aperture are obtained through the wavefront sensor, the slope is calculated and a slope vector is formed, the noise covariance matrix is ​​calculated, a small displacement command is applied to each actuator of the deformable mirror, and the slope increment is recorded to form an interaction matrix. Step S402: Based on the interaction matrix and the noise covariance matrix, the control matrix and phase reconstruction matrix are solved by weighted least squares with regularization. Step S403: Under the illumination of the working light, after preprocessing the wavefront sensor image, obtain the current centroid coordinates and calculate the slope, subtract the static bias, and obtain the current slope vector. Step S404: Obtain the correction amount by mapping the current slope vector according to the control matrix, update the deformable mirror command through integral control and leakage, and satisfy the actuator stroke constraint through amplitude limiting projection; Step S405: Calculate the residual slope between the current slope vector and the predicted slope, convert it into residual phase through the phase reconstruction matrix, and obtain the point spread function by combining the aperture function to evaluate the imaging quality; Step S406: Calculate the image quality index based on the residual phase. When the preset threshold is reached or the residual meets the requirements, freeze the deformable mirror command and output the correction result.

[0015] The beneficial effects of this invention are as follows: By accurately constructing the noise covariance matrix and interaction matrix under reference light, and combining it with a weighted least squares method with double regularization, the invention effectively suppresses the influence of measurement noise while balancing actuator travel constraints, significantly improving the robustness of the calibration model. Through the combination of integral control, leakage handling, and amplitude limiting design, the invention avoids actuator saturation due to command accumulation, prevents the superposition of calibration errors, and ensures the stability of the calibration process. The point spread function is obtained through the residual phase and aperture function, intuitively reflecting the imaging quality. Furthermore, the acceptance criteria are quantified using both image quality indicators and the L2 norm of the residual slope vector, freezing deformable mirror commands to ensure the reproducibility of the calibration effect. Overall, this invention meets the requirements of high-precision scenarios such as laser imaging and ranging, providing a stable and reliable optical system state for subsequent tasks. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of a laser wavefront distortion correction system based on adaptive optics according to the present invention; Figure 2 This is a flowchart of a laser wavefront distortion correction method based on adaptive optics according to the present invention.

[0017] In the diagram: Module 101, Module 202, Module 303, Module 404, Module 505, Module 606. Detailed Implementation

[0018] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and implement the subject matter described herein, and changes may be made to the function and arrangement of the elements discussed without departing from the scope of this specification. Various processes or components may be omitted, substituted, or added as needed in the examples. Furthermore, features described in some examples may be combined in other examples.

[0019] It should be noted that, unless otherwise defined, the technical or scientific terms used in one or more embodiments of the present invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in one or more embodiments of the present invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" indicate that the element or object preceding the term encompasses the elements or objects listed following the term and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0020] like Figures 1-2 As shown, a laser wavefront distortion correction system based on adaptive optics includes: The first module 101 is used to obtain the reference centroid coordinates and current centroid coordinates of each sub-aperture through the wavefront sensor under reference light illumination, convert the slope to form a slope vector, calculate the noise covariance matrix, apply small displacement commands to each actuator of the deformable mirror, and record the slope increment to form an interaction matrix. The second module 102 is used to solve the control matrix and phase reconstruction matrix by regularized weighted least squares based on the interaction matrix and the noise covariance matrix. The third module 103 is used to obtain the current centroid coordinates and calculate the slope after preprocessing the wavefront sensor image under working light illumination, and to obtain the current slope vector by subtracting the static bias. The fourth module 104 is used to obtain the correction amount by mapping the current slope vector according to the control matrix, update the deformable mirror command through integral control and leakage, and satisfy the actuator stroke constraint through amplitude limiting projection; The fifth module 105 is used to calculate the residual slope between the current slope vector and the predicted slope, which is then converted into the residual phase by the phase reconstruction matrix and combined with the aperture function to obtain the point spread function to evaluate the imaging quality. The sixth module 106 is used to calculate the image quality index based on the residual phase. When the preset threshold is reached or the residual meets the requirements, the deformable mirror command is frozen and the correction result is output.

[0021] In one embodiment of the present invention, under reference light illumination, for each sub-aperture, the reference pixel intensity of each pixel within that sub-aperture is obtained. The x-coordinate of each pixel is multiplied by the corresponding reference pixel intensity, and the sum is obtained to get a first numerator. The sum of all reference pixel intensities within that sub-aperture is obtained to get a first denominator. The first numerator is divided by the first denominator to obtain the reference x-direction centroid coordinates of that sub-aperture. The y-coordinate of each pixel is multiplied by the corresponding reference pixel intensity, and the sum is obtained to get a second numerator. The second numerator is divided by the first denominator to obtain the reference y-direction centroid coordinates of that sub-aperture. The same method is followed. Calculate the current centroid coordinates of each sub-aperture; subtract the reference x-axis centroid coordinate of each sub-aperture from its current x-axis centroid coordinate, and divide the difference by the focal length of the microlens to obtain the x-axis slope of that sub-aperture; subtract the reference y-axis centroid coordinate of each sub-aperture from its current y-axis centroid coordinate, and divide the difference by the focal length of the microlens to obtain the y-axis slope of that sub-aperture; arrange the x-axis slopes of all sub-apertures in a preset order, and then arrange the y-axis slopes of all sub-apertures in the same preset order to form a slope vector.

[0022] It should be noted that the sub-aperture is a tiny aperture unit in the wavefront sensor that divides the incident light beam, preferably 1024 units; the reference pixel intensity is the light signal intensity value received by each pixel within the sub-aperture under reference light illumination; the current pixel intensity is the light signal intensity value received by each pixel within the same sub-aperture during the measurement phase; the reference x-direction centroid coordinate is the coordinate of the center position of the light beam energy within the sub-aperture in the x-direction under reference light illumination; the reference y-direction centroid coordinate is the coordinate of the center position of the light beam energy within the sub-aperture in the y-direction under reference light illumination; the current x-direction centroid coordinate is the coordinate of the center position of the light beam energy within the sub-aperture in the x-direction during the measurement phase; the current y-direction centroid coordinate is the coordinate of the light beam energy within the sub-aperture during the measurement phase. The coordinates of the center position in the y-direction; the focal length of the microlens is the distance from the focal point of the microlens in the wavefront sensor to the center of the lens, preferably 300 micrometers to 500 micrometers; the slope in the x-direction is the rate of change of the beam wavefront in the x-direction within the sub-aperture; the slope in the y-direction is the rate of change of the beam wavefront in the y-direction within the sub-aperture; the slope vector is a vector composed of the x-direction slopes and y-direction slopes of all sub-apertures arranged in a preset order; the spatial position information of the beam is converted into quantized slope data that can be used for subsequent wavefront correction; the beneficial effect is to provide basic data support for the accurate quantification of wavefront distortion, ensure that subsequent correction calculations can respond specifically to the beam distortion state, and provide a reliable input basis for the entire adaptive optics correction system.

[0023] In one embodiment of the present invention, under the condition that the reference light remains static, slope vectors are acquired multiple times, and the average value of the slopes at corresponding positions in all groups of slope vectors is calculated to obtain the mean value of the slope vectors; the mean value of the slope vectors at corresponding positions is subtracted from the slope at each position in each group of slope vectors to obtain the deviation vector corresponding to each group of slope vectors; each group of deviation vectors is multiplied by its own transpose matrix to obtain the deviation product matrix corresponding to each group; the average value of all groups of deviation product matrices is calculated to obtain the noise covariance matrix; for each actuator of the deformable mirror, an application is applied to the actuator. A preset micro-displacement command is used; the slope vector before applying the micro-displacement command is recorded as the pre-application slope vector; the slope vector after applying the micro-displacement command is recorded as the post-application slope vector; the slope at each position in the post-application slope vector is subtracted from the slope at the corresponding position in the pre-application slope vector to obtain the slope increment corresponding to the actuator; each position increment in the slope increment is divided by the micro-displacement command of the actuator to obtain the matrix column corresponding to the actuator; all matrix columns corresponding to the actuators are arranged sequentially according to the preset numbering order of the actuators to form an interaction matrix.

[0024] It should be noted that the slope vector mean is the average value of the slope at the corresponding position in multiple slope vector acquisitions; the deviation vector is a vector composed of the difference between each group of slope vectors and the slope vector mean; the transpose matrix is ​​the matrix obtained by interchanging the rows and columns of the original matrix; the deviation product matrix is ​​the matrix obtained by multiplying the deviation vector with its transpose matrix; the noise covariance matrix is ​​a matrix that reflects the noise intensity and inter-correlation of each slope measurement value; the deformable mirror is an optical element that achieves wavefront distortion correction by changing the shape of the mirror surface; the actuator is a functional unit on the deformable mirror that drives the mirror surface to produce displacement, preferably 140 in number; the micro-displacement command is a command signal that controls the actuator to produce micro-deformation, and the displacement amount is preferably one-tenth to one-twentieth of the working wavelength, and the displacement amount is 77 nanometers to 155 nanometers when the working wavelength is 1550 nanometers; The slope vector before application is the slope vector acquired before the actuator applies a small displacement command; the slope vector after application is the slope vector acquired after the actuator applies a small displacement command; the slope increment is a vector composed of the difference between the slope vector after application and the slope vector before application; the matrix column is a column vector obtained by dividing the slope increment by the small displacement command; the preset numbering order is a pre-defined sorting method for all actuators of the deformable mirror; the interaction matrix is ​​a matrix formed by arranging the matrix columns corresponding to all actuators in a preset numbering order; the noise characteristics of the slope measurement are quantified, and a mapping relationship between the displacement of the deformable mirror actuator and the slope change is established; the beneficial effect is to provide an accurate noise reference and actuator control model for subsequent closed-loop control, reduce the impact of noise on the correction effect, improve the accuracy and stability of wavefront correction, and ensure that the adaptive optics system can respond efficiently to wavefront distortion.

[0025] In one embodiment of the present invention, the control matrix is ​​calculated using a regularized weighted least squares method, including the following steps: Step S201: Calculate the product of the transpose of the interaction matrix and the inverse of the noise covariance matrix; Step S202: Multiply the product by the interaction matrix to obtain the first intermediate calculation result; Step S203: Calculate the product of the regularization coefficient of the control matrix and the identity matrix to obtain the second intermediate calculation result; Step S204: Add the first intermediate calculation result to the second intermediate calculation result to obtain the third intermediate calculation result; Step S205: Calculate the inverse result of the third intermediate calculation result, and multiply the inverse result with the transpose of the interaction matrix and the inverse of the noise covariance matrix in sequence to obtain the control matrix. The control matrix converts the slope vector into deformable mirror commands.

[0026] It should be noted that the regularization coefficient of the control matrix is ​​a positive number used to suppress high-gain or high-frequency modes, with a preferred value between 0.001 and 0.01; the identity matrix is ​​a square matrix with 1s on the main diagonal and 0s elsewhere, and its dimension is consistent with the number of deformable mirror actuators; the control matrix is ​​a linear mapping matrix that converts the slope vector into deformable mirror commands; the deformable mirror commands are a set of instructions used to drive the actuators of the deformable mirror to produce displacements; by obtaining the linear mapping matrix from the slope vector to the deformable mirror commands, and combining regularization and weighting, the influence of measurement noise and the travel constraints of the deformable mirror are balanced, thereby improving the robustness and real-time performance of the closed-loop control of the adaptive optics system.

[0027] In one embodiment of the present invention, the phase reconstruction matrix is ​​calculated using a regularized weighted least squares method, including the following steps: Step S301: Calculate the product of the transpose of the gradient operator and the inverse of the noise covariance matrix; Step S302: Multiply the product obtained in the first step with the gradient operator to obtain the fourth intermediate calculation result; Step S303: Calculate the product of the transpose of the smoothing prior operator and the smoothing prior operator; Step S304: Calculate the product of the regularization coefficient of the phase reconstruction matrix and the product obtained in the third step to obtain the fifth intermediate calculation result; Step S305: Add the fourth intermediate calculation result to the fifth intermediate calculation result to obtain the sixth intermediate calculation result; Step S306: Calculate the inverse of the sixth intermediate calculation result, and multiply the inverse result by the transpose of the gradient operator and the inverse of the noise covariance matrix in sequence to obtain the phase reconstruction matrix.

[0028] It should be noted that the gradient operator is a discrete operator that converts the aperture phase grid into slope sampling in the corresponding x and y directions, with the sampling density matching the sub-aperture distribution of the wavefront sensor; the smoothing prior operator is a discrete operator used to suppress noise amplification during phase reconstruction, preferably the discrete Laplacian operator; the aperture phase is a physical quantity characterizing laser wavefront distortion, measured in radians, covering the entire optical aperture region; by obtaining the linear mapping matrix from the slope vector to the aperture phase, and through weighted least squares combined with smoothing regularization, the influence of measurement noise on phase reconstruction is reduced, resulting in a smooth and accurate aperture phase estimate, providing a reliable basis for subsequent imaging quality assessment.

[0029] In one embodiment of the present invention, for each pixel of each sub-aperture, the difference between the working light pixel intensity of the pixel and the dark field value of the corresponding sub-aperture is calculated, and it is determined whether the difference is greater than or equal to a preset threshold. If it is greater than or equal to the preset threshold, the threshold mask is set to one; if it is less than the preset threshold, the threshold mask is set to zero. The difference is multiplied by the threshold mask to obtain the preprocessed pixel intensity of each pixel of each sub-aperture. For each sub-aperture, calculate the sum of the products of the lateral coordinates of all pixels in that sub-aperture and the corresponding preprocessed pixel intensities, calculate the sum of the intensities of all preprocessed pixels in that sub-aperture, divide the sum of the products by the sum of the intensities to obtain the current centroid lateral coordinates of that sub-aperture; using the same method, calculate the current centroid longitudinal coordinates of that sub-aperture by using the pixel longitudinal coordinates and the corresponding preprocessed pixel intensities. For each sub-aperture, calculate the difference between the current centroid lateral coordinate and the reference centroid lateral coordinate of the sub-aperture, divide the difference by the microlens focal length to obtain the lateral slope of the sub-aperture; calculate the difference between the current centroid longitudinal coordinate and the reference centroid longitudinal coordinate of the sub-aperture, divide the difference by the microlens focal length to obtain the longitudinal slope of the sub-aperture; concatenate all lateral and longitudinal slopes in the order of lateral slopes of all sub-apertures arranged sequentially, followed by longitudinal slopes of all sub-apertures arranged sequentially after the lateral slopes, to form a slope vector without offset subtraction; The current slope vector is obtained by subtracting the slope vector without bias subtraction from the slope vector with static bias.

[0030] It should be noted that: the working light pixel intensity is the light signal intensity value received by the sensor pixel under working light illumination; the threshold mask is a binary identifier used to filter effective pixels, with a value of only one or zero; the preset threshold is a critical value used to distinguish effective signals from noise, preferably a grayscale value between 10 and 50; the preprocessed pixel intensity is the effective pixel intensity value after dark subtraction and threshold filtering; the current centroid horizontal coordinate is the weighted average of the horizontal positions of effective pixels within the sub-aperture at the current moment; the current centroid vertical coordinate is the weighted average of the vertical positions of effective pixels within the sub-aperture at the current moment; and the reference centroid horizontal coordinate is the sub-aperture reference horizontal centroid position determined during calibration. The reference centroid longitudinal coordinate is the position of the sub-aperture reference longitudinal centroid determined during calibration; the transverse slope is the rate of change of the wavefront in the transverse direction within the sub-aperture; the longitudinal slope is the rate of change of the wavefront in the longitudinal direction within the sub-aperture; the slope vector without bias subtraction is a vector formed by arranging the transverse and longitudinal slopes of all sub-apertures in a fixed order; the current slope vector is the slope data set used for subsequent control after subtracting the static bias; the optical signal collected by the sensor is converted into a slope vector reflecting the wavefront distortion, and through processing such as dark field subtraction, threshold filtering, and bias removal, the influence of noise and fixed errors is reduced to obtain accurate wavefront slope information, providing reliable input for closed-loop control.

[0031] In one embodiment of the present invention, the integral gain is multiplied by the control matrix, the resulting product is multiplied by the current slope vector, and the result is negative to obtain the correction vector. Subtract the leakage coefficient from 1, multiply the difference by the current deformable mirror command vector, and add the product to the correction vector to obtain the pre-updated deformable mirror command vector. For each component of the pre-updated deformable mirror command vector, compare the component with the corresponding component of the lower bound vector of the executor command and the corresponding component of the upper bound vector of the executor command. If the component is less than the corresponding component of the lower bound vector of the executor command, then take the corresponding component of the lower bound vector of the executor command as the component at that position. If the component is greater than the corresponding component of the upper bound vector of the executor command, then take the corresponding component of the upper bound vector of the executor command as the component at that position. If the component is between the corresponding components of the lower bound vector of the executor command and the corresponding components of the upper bound vector of the executor command, then keep the component unchanged to obtain the bounded deformable mirror command vector.

[0032] It should be noted that the integral gain is a parameter used to adjust the control response speed, with a preferred value between 0.1 and 1; the leakage coefficient is a parameter used to attenuate historical commands to avoid integral saturation, with a preferred value between 0.001 and 0.01; the correction vector is the deformable mirror command adjustment amount calculated based on the current slope; the pre-updated deformable mirror command vector is the command vector that has undergone integral control and leakage processing but has not been limited; the limited deformable mirror command vector is the final deformable mirror drive command that satisfies the actuator stroke constraint; the actuator command lower limit vector is a vector composed of the minimum drive command value allowed by each actuator; the actuator command upper limit vector is a vector composed of the maximum drive command value allowed by each actuator; by converting the current slope vector reflecting wavefront distortion into the deformable mirror drive command, the correction amount is accumulated through integral control to reduce steady-state error, leakage processing avoids saturation caused by long-term command accumulation, and limiting ensures that the command is within the actuator stroke range, thereby achieving stable and reliable wavefront correction.

[0033] In one embodiment of the present invention, the interaction matrix is ​​multiplied with the current deformable mirror command vector to obtain the predicted slope vector; the current slope vector is subtracted from the predicted slope vector to obtain the residual slope vector; the phase reconstruction matrix is ​​multiplied with the residual slope vector to obtain the residual phase; the imaginary unit is multiplied with the residual phase, the product is exponentially multiplied, and the result is multiplied with the aperture function to obtain the aperture complex amplitude; a Fourier transform operation is performed on the aperture complex amplitude, the modulus of the Fourier transform result is taken, and the result after taking the modulus is squared to obtain the point spread function.

[0034] It should be noted that the predicted slope vector is the theoretical slope vector calculated based on the interaction matrix and the current deformable mirror command vector; the residual slope vector is the difference vector between the current measured slope vector and the predicted slope vector; the residual phase is a physical quantity characterizing the uncorrected wavefront distortion; the aperture complex amplitude is a complex physical quantity that integrates the aperture light transmission characteristics and the residual phase; the imaginary unit is a mathematical identifier used to describe the phase part of the complex amplitude, and its value satisfies the condition that the square equals negative one; the Fourier transform operation is a mathematical operation that converts the complex amplitude in the aperture domain into a signal in the image plane domain; the modulus operation is a mathematical operation that calculates the modulus, i.e., the absolute value, of a complex number; the squaring operation is a mathematical operation that multiplies a value by itself; and the point spread function is a function that describes the imaging effect of the optical system on a point light source. By obtaining the residual information through the difference between the measured slope and the predicted slope, the point spread function is obtained through phase reconstruction and a series of mathematical operations, which intuitively reflects the imaging effect after wavefront correction and provides a direct basis for subsequent imaging quality assessment.

[0035] In one embodiment of the present invention, the residual phases corresponding to all sampling points in the aperture sampling point set are summed, and the sum is divided by the number of aperture sampling points to obtain the average value of the residual phases. For the residual phase of each sampling point in the aperture sampling point set, calculate the difference between the residual phase and the average residual phase, square each difference, sum all the squared results, divide the sum by the number of aperture sampling points, and take the square root of the quotient to obtain the root mean square error of the residual wavefront. Calculate the ratio of 2π to the working wavelength, square the ratio, calculate the square of the root mean square of the residual wavefront error, multiply the results of the two square operations, take the negative value of the product, and perform an exponential operation on the negative value to obtain the image quality index. Squaring each component of the residual slope vector, summing all the squaring results, and taking the square root of the sum to obtain the L2 norm of the residual slope vector. If the image quality index is greater than or equal to the image quality index threshold, or the L2 norm of the residual slope vector is less than or equal to the residual slope vector threshold, then the stopping condition is met. When the stopping condition is met, the current deformable mirror command vector is determined as the frozen deformable mirror command vector, and the frozen deformable mirror command vector and the corresponding correction result are output.

[0036] It should be noted that the aperture sampling point set is a set of discrete sampling positions covering the entire optical aperture; the number of aperture sampling points is the total number of sampling positions included in the aperture sampling point set, preferably between 1024 and 4096; the average residual phase is the arithmetic mean of the residual phases corresponding to all aperture sampling points; the root mean square of the residual wavefront error is a scalar parameter characterizing the overall fluctuation of the residual phase; the image quality index is a parameter that quantifies the image sharpness based on the root mean square of the residual wavefront error, with a value between 0 and 1; the image quality index threshold is used to judge... The critical value for whether the imaging quality meets the standard is preferably 0.8; the L2 norm of the residual slope vector is a mathematical quantity that characterizes the overall magnitude of the residual slope vector; the residual slope vector threshold is a critical value for judging whether the residual slope is small enough, preferably between 0.01 and 0.1; the frozen deformable mirror command vector is the final deformable mirror drive command determined when the stopping condition is met; the stopping condition is the criterion for determining whether to terminate the wavefront correction closed-loop process; the correction result is the output data containing the frozen deformable mirror command vector and the corresponding imaging quality parameters.

[0037] It should be noted that the freeze deformable mirror command maintains the stability of the correction effect, avoids mirror fluctuations caused by continuous closed-loop adjustment, and ensures that the image quality is always at the standard level. The image quality index is obtained by quantifying the residual wavefront error and residual slope. The correction is judged according to the dual judgment criteria. After the standard is met, the freeze deformable mirror command is executed and the result is output, ensuring that the correction effect can be quantified and accepted, and providing a stable optical system state for subsequent imaging applications.

[0038] In one embodiment of the present invention, such as Figure 2 As shown, a laser wavefront distortion correction method based on adaptive optics includes the following steps: Step S401: Under the illumination of the reference light, the reference centroid coordinates and the current centroid coordinates of each sub-aperture are obtained through the wavefront sensor, the slope is calculated and a slope vector is formed, the noise covariance matrix is ​​calculated, a small displacement command is applied to each actuator of the deformable mirror, and the slope increment is recorded to form an interaction matrix. Step S402: Based on the interaction matrix and the noise covariance matrix, the control matrix and phase reconstruction matrix are solved by weighted least squares with regularization. Step S403: Under the illumination of the working light, after preprocessing the wavefront sensor image, obtain the current centroid coordinates and calculate the slope, subtract the static bias, and obtain the current slope vector. Step S404: Obtain the correction amount by mapping the current slope vector according to the control matrix, update the deformable mirror command through integral control and leakage, and satisfy the actuator stroke constraint through amplitude limiting projection; Step S405: Calculate the residual slope between the current slope vector and the predicted slope, convert it into residual phase through the phase reconstruction matrix, and obtain the point spread function by combining the aperture function to evaluate the imaging quality; Step S406: Calculate the image quality index based on the residual phase. When the preset threshold is reached or the residual meets the requirements, freeze the deformable mirror command and output the correction result.

[0039] It should be noted that the interval and threshold sizes are set for ease of comparison. The size of the threshold depends on the amount of sample data and the base number set by those skilled in the art for each set of sample data, as long as it does not affect the proportional relationship between the parameter and the quantized value. Furthermore, the above formulas are all dimensionless calculations, and the formulas are derived from software simulations using a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0040] The embodiments of this example have been described above. However, this example is not limited to the specific implementation methods described above. The specific implementation methods described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms based on the guidance of this example, and all of them are within the protection scope of this example.

Claims

1. A laser wavefront distortion correction system based on adaptive optics, characterized in that, include: The first module is used to obtain the reference centroid coordinates and current centroid coordinates of each sub-aperture through a wavefront sensor under reference light illumination, convert the slope to form a slope vector, calculate the noise covariance matrix, apply small displacement commands to each actuator of the deformable mirror, and record the slope increment to form an interaction matrix. The second module is used to solve the control matrix and phase reconstruction matrix using regularized weighted least squares based on the interaction matrix and the noise covariance matrix. The third module is used to obtain the current centroid coordinates and calculate the slope after preprocessing the wavefront sensor image under working light illumination, and to obtain the current slope vector by subtracting the static bias. The fourth module is used to obtain the correction amount by mapping the current slope vector according to the control matrix, update the deformable mirror command through integral control and leakage, and satisfy the actuator stroke constraint through amplitude limiting projection. The fifth module is used to calculate the residual slope between the current slope vector and the predicted slope, which is then converted into the residual phase by the phase reconstruction matrix. Combined with the aperture function, the point spread function is obtained to evaluate the imaging quality. The sixth module is used to calculate the image quality index based on the residual phase. When the preset threshold is reached or the residual meets the requirements, the deformable mirror command is frozen and the correction result is output.

2. The laser wavefront distortion correction system based on adaptive optics according to claim 1, characterized in that, Illuminated by reference light, for each sub-aperture, the reference pixel intensity of each pixel within that sub-aperture is obtained. The x-coordinate of each pixel is multiplied by the corresponding reference pixel intensity, and the sum is obtained to get the first numerator. The sum of the intensities of all reference pixels within that sub-aperture is obtained to get the first denominator. The first numerator is divided by the first denominator to obtain the reference x-direction centroid coordinates of that sub-aperture. The y-coordinate of each pixel is multiplied by the corresponding reference pixel intensity, and the sum is obtained to get the second numerator. The second numerator is divided by the first denominator to obtain the reference y-direction centroid coordinates of that sub-aperture. The same method is used to calculate the centroid coordinates of each sub-aperture. The current centroid coordinates of the aperture; subtract the reference centroid coordinates of the sub-aperture from the current x-axis centroid coordinates of each sub-aperture, and divide the difference by the focal length of the microlens to obtain the x-axis slope of that sub-aperture; subtract the reference y-axis centroid coordinates of the sub-aperture from the current y-axis centroid coordinates of each sub-aperture, and divide the difference by the focal length of the microlens to obtain the y-axis slope of that sub-aperture; arrange the x-axis slopes of all sub-apertures in a preset order, and then arrange the y-axis slopes of all sub-apertures in the same preset order to form a slope vector.

3. The laser wavefront distortion correction system based on adaptive optics according to claim 2, characterized in that, With the reference light remaining static, slope vectors are acquired multiple times, and the average slope at corresponding positions in all groups of slope vectors is calculated to obtain the mean slope vector. The mean slope vector is obtained by subtracting the mean slope vector at the corresponding position from the slope at each position in each group of slope vectors. The deviation vector for each group of slope vectors is then multiplied by its own transpose to obtain the deviation product matrix for each group. The average deviation product matrix for all groups is calculated to obtain the noise covariance matrix. For each actuator of the deformable mirror, a preset small displacement command is applied to the actuator. Record the slope vector before applying the small displacement command as the pre-application slope vector; record the slope vector after applying the small displacement command as the post-application slope vector; subtract the slope of the corresponding position in the pre-application slope vector from the slope of each position in the post-application slope vector to obtain the slope increment corresponding to the actuator; divide each position increment in the slope increment by the small displacement command of the actuator to obtain the matrix column corresponding to the actuator; arrange the matrix columns corresponding to all actuators in the preset numbering order of the actuators to form an interaction matrix.

4. The laser wavefront distortion correction system based on adaptive optics according to claim 1, characterized in that, The control matrix is ​​calculated using a weighted least squares method with regularization, including the following steps: Step S201: Calculate the product of the transpose of the interaction matrix and the inverse of the noise covariance matrix; Step S202: Multiply the product by the interaction matrix to obtain the first intermediate calculation result; Step S203: Calculate the product of the regularization coefficient of the control matrix and the identity matrix to obtain the second intermediate calculation result; Step S204: Add the first intermediate calculation result to the second intermediate calculation result to obtain the third intermediate calculation result; Step S205: Calculate the inverse result of the third intermediate calculation result, and multiply the inverse result with the transpose of the interaction matrix and the inverse of the noise covariance matrix in sequence to obtain the control matrix. The control matrix converts the slope vector into deformable mirror commands.

5. A laser wavefront distortion correction system based on adaptive optics according to claim 4, characterized in that, The phase reconstruction matrix is ​​calculated using a weighted least squares method with regularization, including the following steps: Step S301: Calculate the product of the transpose of the gradient operator and the inverse of the noise covariance matrix; Step S302: Multiply the product obtained in the first step with the gradient operator to obtain the fourth intermediate calculation result; Step S303: Calculate the product of the transpose of the smoothing prior operator and the smoothing prior operator; Step S304: Calculate the product of the regularization coefficient of the phase reconstruction matrix and the product obtained in the third step to obtain the fifth intermediate calculation result; Step S305: Add the fourth intermediate calculation result to the fifth intermediate calculation result to obtain the sixth intermediate calculation result; Step S306: Calculate the inverse of the sixth intermediate calculation result, and multiply the inverse result by the transpose of the gradient operator and the inverse of the noise covariance matrix in sequence to obtain the phase reconstruction matrix.

6. The laser wavefront distortion correction system based on adaptive optics according to claim 1, characterized in that, For each pixel of each sub-aperture, calculate the difference between the working light pixel intensity of that pixel and the dark field value of the corresponding sub-aperture, and determine whether the difference is greater than or equal to a preset threshold. If it is greater than or equal to the preset threshold, the threshold mask is set to one; if it is less than the preset threshold, the threshold mask is set to zero. Multiply the above difference by the threshold mask to obtain the preprocessed pixel intensity of each pixel of each sub-aperture. For each sub-aperture, calculate the sum of the products of the lateral coordinates of all pixels in that sub-aperture and the corresponding preprocessed pixel intensities, calculate the sum of the intensities of all preprocessed pixels in that sub-aperture, divide the sum of the products by the sum of the intensities to obtain the current centroid lateral coordinates of that sub-aperture; using the same method, calculate the current centroid longitudinal coordinates of that sub-aperture by using the pixel longitudinal coordinates and the corresponding preprocessed pixel intensities. For each sub-aperture, calculate the difference between the current centroid lateral coordinate and the reference centroid lateral coordinate of the sub-aperture, divide the difference by the microlens focal length to obtain the lateral slope of the sub-aperture; calculate the difference between the current centroid longitudinal coordinate and the reference centroid longitudinal coordinate of the sub-aperture, divide the difference by the microlens focal length to obtain the longitudinal slope of the sub-aperture; concatenate all lateral and longitudinal slopes in the order of lateral slopes of all sub-apertures arranged sequentially, followed by longitudinal slopes of all sub-apertures arranged sequentially after the lateral slopes, to form a slope vector without offset subtraction; The current slope vector is obtained by subtracting the slope vector without bias subtraction from the slope vector with static bias.

7. A laser wavefront distortion correction system based on adaptive optics according to claim 1, characterized in that, Multiply the integral gain by the control matrix, multiply the product by the current slope vector, and take the negative value of the result to obtain the correction vector; Subtract the leakage coefficient from 1, multiply the difference by the current deformable mirror command vector, and add the product to the correction vector to obtain the pre-updated deformable mirror command vector. For each component of the pre-updated deformable mirror command vector, compare the component with the corresponding component of the lower bound vector of the executor command and the corresponding component of the upper bound vector of the executor command. If the component is less than the corresponding component of the lower bound vector of the executor command, then take the corresponding component of the lower bound vector of the executor command as the component at that position. If the component is greater than the corresponding component of the upper bound vector of the executor command, then take the corresponding component of the upper bound vector of the executor command as the component at that position. If the component is between the corresponding components of the lower bound vector of the executor command and the corresponding components of the upper bound vector of the executor command, then keep the component unchanged to obtain the bounded deformable mirror command vector.

8. The laser wavefront distortion correction system based on adaptive optics according to claim 1, characterized in that, The interaction matrix is ​​multiplied by the current deformable mirror command vector to obtain the predicted slope vector; the current slope vector is subtracted from the predicted slope vector to obtain the residual slope vector; the phase reconstruction matrix is ​​multiplied by the residual slope vector to obtain the residual phase; the imaginary unit is multiplied by the residual phase, the product is exponentially multiplied, and the result is multiplied by the aperture function to obtain the aperture complex amplitude; a Fourier transform is performed on the aperture complex amplitude, the modulus of the Fourier transform result is taken, and the result is squared to obtain the point spread function.

9. A laser wavefront distortion correction system based on adaptive optics according to claim 1, characterized in that, The residual phases corresponding to all sampling points in the aperture sampling point set are summed, and the sum is divided by the number of aperture sampling points to obtain the average value of the residual phases. For the residual phase of each sampling point in the aperture sampling point set, calculate the difference between the residual phase and the average residual phase, square each difference, sum all the squared results, divide the sum by the number of aperture sampling points, and take the square root of the quotient to obtain the root mean square error of the residual wavefront. Calculate the ratio of 2π to the working wavelength, square the ratio, calculate the square of the root mean square of the residual wavefront error, multiply the results of the two square operations, take the negative value of the product, and perform an exponential operation on the negative value to obtain the image quality index. Squaring each component of the residual slope vector, summing all the squaring results, and taking the square root of the sum to obtain the L2 norm of the residual slope vector. If the image quality index is greater than or equal to the image quality index threshold, or the L2 norm of the residual slope vector is less than or equal to the residual slope vector threshold, then the stopping condition is met. When the stopping condition is met, the current deformable mirror command vector is determined as the frozen deformable mirror command vector, and the frozen deformable mirror command vector and the corresponding correction result are output.

10. A laser wavefront distortion correction method based on adaptive optics, characterized in that, Implementing a laser wavefront distortion correction system based on adaptive optics as described in any one of claims 1 to 9 includes the following steps: Step S401: Under the illumination of the reference light, the reference centroid coordinates and the current centroid coordinates of each sub-aperture are obtained through the wavefront sensor, the slope is calculated and a slope vector is formed, the noise covariance matrix is ​​calculated, a small displacement command is applied to each actuator of the deformable mirror, and the slope increment is recorded to form an interaction matrix. Step S402: Based on the interaction matrix and the noise covariance matrix, the control matrix and phase reconstruction matrix are solved by weighted least squares with regularization. Step S403: Under the illumination of the working light, after preprocessing the wavefront sensor image, obtain the current centroid coordinates and calculate the slope, subtract the static bias, and obtain the current slope vector. Step S404: Obtain the correction amount by mapping the current slope vector according to the control matrix, update the deformable mirror command through integral control and leakage, and satisfy the actuator stroke constraint through amplitude limiting projection; Step S405: Calculate the residual slope between the current slope vector and the predicted slope, convert it into residual phase through the phase reconstruction matrix, and obtain the point spread function by combining the aperture function to evaluate the imaging quality; Step S406: Calculate the image quality index based on the residual phase. When the preset threshold is reached or the residual meets the requirements, freeze the deformable mirror command and output the correction result.

Images