Mirror convergence method and device based on double-feedback surface shape regulation and control optimization

Abstract

The invention provides a mirror convergence method and device based on double-feedback surface shape regulation and control optimization. The method comprises the following steps: monitoring wavefront components modulated by a deformed mirror surface shape in real time by using a wavefront sensor; judging whether the surface shape coefficient residual error of the wavefront component is greater than or equal to a set threshold value; if yes, the surface shape reconstruction voltage is rapidly solved through the influence function matrix, and the deformable mirror is rapidly converged based on the surface shape reconstruction voltage; if yes, using an image sensor to monitor the change of light intensity distribution or imaging definition, and using the change as an evaluation parameter of the distortion wavefront correction degree; and based on the evaluation parameters, the surface shape reconstruction precision is improved through a search optimization algorithm. According to the method, the defects that the control precision of a closed-loop control algorithm based on an influence function is influenced by the measurement precision of an influence function matrix, and a closed-loop control algorithm without an influence function is low in calculation efficiency and easy to converge to a local optimal value are overcome, and the requirements of a fixed zoom image stabilization system can be met.

Description

Technical Field

[0001] The present invention relates to the technical field of surface shape control and optimization, and in particular to a mirror convergence method and device based on double-feedback surface shape control and optimization. Background Art

[0002] Deformable mirror is an important wavefront corrector in the field of adaptive optics. A typical adaptive optical system is mainly composed of a wavefront sensor, a wavefront controller, a wavefront corrector and some auxiliary optical elements. Figure 1 As shown in the dashed box in the middle. The wavefront sensor detects and displays the aberration composition of the distorted wavefront in real time, and feeds the detection results back to the wavefront controller, which solves the measured wavefront information and solves the control voltage signal. The wavefront corrector receives the control voltage signal output from the wavefront controller and performs conjugate compensation for the system wavefront distortion.

[0003] The piezoelectric stacked deformable mirror has the advantages of simple mechanical structure, mature research and development technology, and great actuation potential. It is a wavefront corrector with high adaptability and mature manufacturing process. The inverse piezoelectric effect of piezoelectric ceramics allows the material to produce mechanical deformation under the action of an electric field. After the voltage is applied, each piezoelectric sheet will produce deformation in the same direction, which will gradually accumulate to the actuator terminal, thereby driving the deformable mirror surface to produce corresponding deformation.

[0004] As a high-precision wavefront controller, the surface control accuracy of the deformable mirror is of great significance to the overall working performance of the device. Deformable mirrors often have dozens or even thousands of actuation channels and have great deformation freedom. Factors such as model errors, structural cross-linking coupling and piezoelectric actuation hysteresis will introduce control errors. Precise control of the mirror surface shape is a difficult problem in the application of deformable mirrors.

[0005] The current deformable mirror convergence algorithm can be roughly divided into two categories: open-loop control and closed-loop control, depending on whether there is a closed-loop feedback link. The closed-loop algorithm can be divided into closed-loop control algorithms based on influence function theory and without influence function, depending on whether it is necessary to solve the influence function matrix in advance to control the voltage iteration. The open-loop control algorithm only needs to perform a single surface shape or wavefront measurement, and the solution efficiency is high. It can achieve fast response reconstruction of the surface shape, but the deformable mirror itself does not have the correction capability, and the accuracy and robustness are poor. Therefore, the use of only open-loop algorithms cannot adapt to the field with high requirements for surface shape accuracy.

[0006] In order to improve the accuracy of deformable mirror reconstruction, voltage iteration can be performed by introducing sensor feedback information. When wavefront sensors, which directly sense wavefront or surface components, are used, surface iteration depends on the measurement of the influence function matrix, which is called a closed-loop control algorithm based on influence function theory. It can improve the robustness of the control process, but the convergence speed is slow, and the construction process of the influence function matrix does not consider the influence of structural cross-linking coupling, resulting in a high error level. This type of algorithm has a stable iterative convergence process and better control accuracy than open loop, but the accuracy of the final converged surface shape is greatly affected by the accuracy of the influence function matrix.

[0007] When an image sensor is selected that indirectly infers the wavefront or surface components through image processing, the surface shape iteration process does not need to affect the function matrix, which is called a closed-loop control algorithm with no influence function. The accuracy of this type of algorithm is not affected by the error in solving the influence function matrix, but it often requires thousands of search optimizations to converge, and there is a possibility of converging to a local optimal value, which limits its scope of application.

[0008] From the analysis, it can be seen that the closed-loop control algorithm based on the influence function has good control efficiency, but its control accuracy is affected by the measurement accuracy of the influence function matrix, and the surface reconstruction accuracy cannot be further improved. The closed-loop control algorithm without the influence function can generally achieve a higher precision surface shape, but it often requires thousands of iterations to converge, and there is a possibility of converging to a local optimal value. Therefore, the existing algorithms cannot maintain the stability of control accuracy under the condition of variable aperture and variable aberration surface shape, and cannot meet the requirements of the fixed zoom stabilization system. Summary of the invention

[0009] In view of this, the present invention provides a mirror convergence method and device based on dual-feedback surface shape control optimization to solve at least one of the above-mentioned problems.

[0010] In order to achieve the above object, the present invention adopts the following scheme:

[0011] According to a first aspect of the present invention, there is provided a mirror convergence method based on dual-feedback surface shape control optimization, the method comprising: using a wavefront sensor to monitor in real time the wavefront component after deformable mirror surface shape modulation; judging whether the surface shape coefficient residual of the wavefront component is greater than or equal to a set threshold; if the surface shape coefficient residual is greater than or equal to the set threshold, quickly obtaining a surface shape reconstruction voltage through an influence function matrix, and making the deformable mirror converge quickly based on the surface shape reconstruction voltage, and then returning to continue judging whether the surface shape coefficient residual of the wavefront component is greater than or equal to the set threshold; if the surface shape coefficient residual is less than the set threshold, using an image sensor to monitor the change in light intensity distribution or imaging clarity as an evaluation parameter of the degree of distortion wavefront correction; based on the evaluation parameter, improving the surface shape reconstruction accuracy through a small step size and high-order search optimization algorithm so that the surface shape reconstruction accuracy reaches the theoretical optimal value of the search optimization algorithm.

[0012] As an embodiment of the present invention, in the above method, if the surface shape coefficient residual is less than the set threshold, the method also includes: after the system realizes the surface shape of the current focal length once, storing the key-value pair mapping relationship of "focal length-channel voltage" in a hash table to quickly realize system zoom.

[0013] As an embodiment of the present invention, if the surface shape coefficient residual in the above method is less than the set threshold, the method also includes: after the system realizes the surface shape of the current focal length once, storing the key-value pair mapping relationship of "tilt angle-channel voltage" in a hash table to quickly realize system image stabilization.

[0014] As an embodiment of the present invention, before using a wavefront sensor to monitor in real time the wavefront component after the deformable mirror shape is modulated, the method further includes: determining whether the target focal length or target inclination angle is in a hash table, and if so, directly reading the corresponding channel voltage in the hash table, and controlling the deformable mirror based on the channel voltage; if not, using a wavefront sensor to monitor in real time the wavefront component after the deformable mirror shape is modulated.

[0015] As an embodiment of the present invention, the method of quickly obtaining the surface reconstruction voltage by using the influence function matrix includes: using the surface characterization vector difference between the target surface shape and the current surface shape as an evaluation function, and the evaluation function is shown in the following formula:

[0016] J(x)=e n =S d -S n =S d -GC n

[0017] In the above formula, J(x) is the evaluation function, e nis the first four-order surface characterization vector deviation between the target surface and the current surface, S d is the target surface characterization vector, C n is the control voltage signal currently acting on the deformable mirror, S n is the characterization vector of the current deformed mirror shape, and G is the influence function matrix.

[0018] Under the guidance of the gradient descent method, the reconstructed voltage signal is adjusted in the opposite direction of the gradient of the evaluation function, and the reconstructed voltage signal of the deformable mirror in the next iteration is:

[0019] C n+1 =C n +αG T e n

[0020] In the above formula, C n+1 Reconstruct the voltage signal of the deformable mirror for the next iteration, and α is the gain coefficient.

[0021] As an embodiment of the present invention, the search optimization algorithm in the above method is a stochastic parallel gradient descent algorithm, and the evaluation parameters are:

[0022]

[0023] In the above formula, E is the evaluation parameter, g(x,y) represents the grayscale value of the image at (x,y), and g(x+1,y) represents the grayscale value of the image at (x+1,y).

[0024] As an embodiment of the present invention, the wavefront sensor and the image sensor in the above method are implemented using a high-precision interferometer.

[0025] According to a second aspect of the present invention, there is provided a mirror convergence device based on dual-feedback surface shape control optimization, the device comprising: a wavefront monitoring unit, for monitoring the wavefront component after deformed mirror surface shape modulation in real time by using a wavefront sensor; a judgment unit, for judging whether the surface shape coefficient residual of the wavefront component is greater than or equal to a set threshold; a fast reconstruction unit, for quickly obtaining a surface shape reconstruction voltage through an influence function matrix when the judgment unit judges that the surface shape coefficient residual is greater than or equal to the set threshold, and making the deformable mirror converge quickly based on the surface shape reconstruction voltage; an image monitoring unit, for monitoring the change of light intensity distribution or imaging clarity by using an image sensor as an evaluation parameter of the degree of distortion wavefront correction when the judgment unit judges that the surface shape coefficient residual is less than the set threshold; a secondary reconstruction unit, for improving the surface shape reconstruction accuracy through a small step size and high order search optimization algorithm based on the evaluation parameter so that the surface shape reconstruction accuracy reaches the theoretical optimal value of the search optimization algorithm.

[0026] As an embodiment of the present invention, the above-mentioned device also includes: a storage unit, which is used to store the key-value pair mapping relationship of "focal length-channel voltage" in a hash table when the judgment unit judges that the surface coefficient residual is less than the set threshold and after the system has realized the surface shape of the current focal length once, so as to quickly realize system zoom.

[0027] As an embodiment of the present invention, the above-mentioned storage unit is also used to store the key-value pair mapping relationship of "tilt angle-channel voltage" in a hash table when the judgment unit judges that the surface coefficient residual is less than the set threshold and after the system has realized the surface shape of the current focal length once, so as to quickly realize system image stabilization.

[0028] As an embodiment of the present invention, the above-mentioned device also includes: a hash table judgment unit, used to determine whether the target focal length or target inclination angle is in the hash table. If so, directly read the corresponding channel voltage in the hash table, and control the deformable mirror based on the channel voltage; if not, use the wavefront sensor to monitor the wavefront component after the deformable mirror surface shape is modulated in real time.

[0029] As an embodiment of the present invention, the fast reconstruction unit quickly obtains the surface reconstruction voltage through the influence function matrix, including: using the surface characterization vector difference between the target surface shape and the current surface shape as an evaluation function, and the evaluation function is shown in the following formula:

[0030] J(x)=e n =S d -S n =S d -GC n

[0031] In the above formula, J(x) is the evaluation function, e n is the first four-order surface characterization vector deviation between the target surface and the current surface, S d is the target surface characterization vector, C n is the control voltage signal currently acting on the deformable mirror, S n is the characterization vector of the current deformed mirror shape, and G is the influence function matrix.

[0032] Under the guidance of the gradient descent method, the reconstructed voltage signal is adjusted in the opposite direction of the gradient of the evaluation function, and the reconstructed voltage signal of the deformable mirror in the next iteration is:

[0033] C n+1 =C n +αG T e n

[0034] In the above formula, C n+1 Reconstruct the voltage signal of the deformable mirror for the next iteration, and α is the gain coefficient.

[0035] As an embodiment of the present invention, the above search optimization algorithm is a stochastic parallel gradient descent algorithm, and the evaluation parameters are:

[0036]

[0037] In the above formula, E is the evaluation parameter, g(x,y) represents the grayscale value of the image at (x,y), and g(x+1,y) represents the grayscale value of the image at (x+1,y).

[0038] As an embodiment of the present invention, the wavefront sensor and the image sensor in the above device are implemented using a high-precision interferometer.

[0039] According to a third aspect of the present invention, there is provided an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above method when executing the computer program.

[0040] According to a fourth aspect of the present invention, there is provided a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps of the above method when executed by a processor.

[0041] It can be seen from the above technical scheme that the mirror convergence method and device based on dual feedback surface shape control optimization provided by the present invention make up for the defects that the control accuracy of the closed-loop control algorithm based on the influence function is affected by the measurement accuracy of the influence function matrix, and the closed-loop control algorithm without the influence function has low calculation efficiency and is easy to converge to the local optimal value. It can maintain the stability of control accuracy under the condition of variable aperture and variable aberration surface shape, and meet the requirements of the fixed zoom stabilization system. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0043] Figure 1 This is a typical structural diagram of an adaptive optical system;

[0044] Figure 2 It is a flow chart of a mirror convergence method based on dual feedback surface shape control optimization provided in an embodiment of the present application;

[0045] Figure 3 It is a connection diagram of deformable mirror shape control based on ZYGO interferometer provided in an embodiment of the present application;

[0046] Figure 4 It is a schematic diagram of the principle of the deformable mirror shape optimization control solution provided in the embodiment of the present application;

[0047] Figure 5 It is a flow chart of a mirror convergence method in a fixed zoom stabilization system provided by an embodiment of the present application;

[0048] Figure 6 It is a schematic diagram of the reconstruction error of the existing algorithm for the variable aperture and variable aberration surface shape provided by the embodiment of the present application;

[0049] Figure 7 It is a schematic diagram of the reconstruction error of the target surface shape with variable aperture and variable aberration components by the existing control method provided in the embodiment of the present application;

[0050] Figure 8 It is a schematic diagram of the average calculation cost of the existing algorithm provided in the embodiment of the present application;

[0051] Fig. 9 It is a schematic diagram of reconstruction error of the variable aperture and variable aberration surface shape by the dual feedback surface shape control provided in the embodiment of the present application;

[0052] Fig.10 is a schematic diagram of surface reconstruction error statistics provided by an embodiment of the present application;

[0053] Fig.11 It is a schematic diagram comparing the average computational cost of DFCS provided by the embodiment of the present application and the existing algorithm;

[0054] Fig.12 It is a schematic diagram comparing the reconstruction of the surface shape required for zoom stabilization by DFCS and ILC provided in an embodiment of the present application;

[0055] Fig.13 It is a structural schematic diagram of a mirror convergence device based on dual feedback surface shape control optimization provided in an embodiment of the present application;

[0056] Fig.14 It is a structural schematic diagram of a mirror convergence device based on dual-feedback surface shape control optimization provided by another embodiment of the present application;

[0057] Fig.15 It is a structural schematic diagram of a mirror convergence device based on dual-feedback surface shape control optimization provided by another embodiment of the present application;

[0058] Fig.16 It is a schematic block diagram of the system structure of the electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0059] To make the purpose, technical solution and advantages of the embodiments of the present invention more clear, the embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings. Here, the exemplary embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0060] like Figure 2 The figure is a flow chart of a mirror convergence method based on dual feedback surface shape control optimization provided by an embodiment of the present application, the method comprising the following steps:

[0061] Step S101: using a wavefront sensor to monitor in real time the wavefront component modulated by the deformable mirror.

[0062] In this embodiment, the wavefront sensor is used to monitor the wavefront component after the deformable mirror is modulated in real time. It can capture and analyze the phase change and wavefront distortion of the light wave after passing through the deformable mirror. This process is real-time, which means that the wavefront sensor continuously monitors the wavefront changes to ensure that the system can quickly respond to any wavefront distortion.

[0063] Step S102: Determine whether the surface coefficient residual of the wavefront component is greater than or equal to a set threshold. If it is greater than or equal to the set threshold, proceed to step S103; if it is less than the set threshold, proceed to step S104.

[0064] In this embodiment, the surface coefficient refers to a parameter used to describe the current surface state of the deformable mirror, which can be a Zernike coefficient or a surface characterization vector in other forms. The surface coefficient residual refers to the difference between the surface characterization vectors of the target surface shape and the current mirror surface shape, which can be used to quantify the degree of deviation between the current deformable mirror surface shape and the desired target surface shape. For example, it can be the deviation of the first four order surface characterization vectors of the target surface shape and the current mirror surface shape.

[0065] Step S103: quickly obtain the surface reconstruction voltage through the influence function matrix, and make the deformable mirror converge quickly based on the surface reconstruction voltage, and then return to step S102 to continue to determine whether the surface coefficient residual of the wavefront component is greater than or equal to the set threshold.

[0066] The influence function matrix is ​​used to describe the influence of each control channel of the deformable mirror on the change of the surface shape. Through this matrix, the control voltage that needs to be applied can be calculated to achieve the desired surface shape adjustment. In this embodiment, an iterative algorithm based on the influence function theory (such as the gradient descent method) is used to solve the reconstruction voltage (channel reconstruction voltage). The principle of the gradient descent method is that if the evaluation function J(x) is differentiable at a, then the gradient of J(x) decreases fastest in the opposite direction of a. Through this method, the surface shape of the deformable mirror can quickly converge to a better state. This process ensures that the system can quickly adjust the surface shape to approach the target surface shape. After the surface shape converges quickly, the system returns to step S102 and continues to determine whether the residual of the surface shape coefficient of the wavefront component is greater than or equal to the set threshold. If the residual is still greater than or equal to the threshold, the surface shape adjustment continues; otherwise, enter the next step of fine control.

[0067] Preferably, the step of "rapidly obtaining the surface reconstruction voltage by using the influence function matrix" may further include:

[0068] The difference between the target face shape and the current face shape is used as the evaluation function, and the evaluation function is shown in the following formula:

[0069] J(x)=e n =S d -S n =S d -GC n

[0070] In the above formula, J(x) is the evaluation function, e n is the first four-order surface characterization vector deviation between the target surface and the current surface, S d is the target surface characterization vector, C n is the control voltage signal currently acting on the deformable mirror, S n is the characterization vector of the current deformed mirror shape, and G is the influence function matrix.

[0071] Under the guidance of the gradient descent method, the reconstructed voltage signal is adjusted in the opposite direction of the gradient of the evaluation function, and the reconstructed voltage signal of the deformable mirror in the next iteration is:

[0072] C n+1 =C n +αG T e n

[0073] In the above formula, C n+1 Reconstruct the voltage signal of the deformable mirror for the next iteration, and α is the gain coefficient.

[0074] Step S104: using an image sensor to monitor changes in light intensity distribution or imaging clarity as evaluation parameters for the degree of distortion wavefront correction.

[0075] By monitoring changes in light intensity distribution or imaging clarity, the image sensor can provide a quantitative evaluation of the current wavefront correction level. This information can be used to determine the effectiveness of wavefront distortion correction. The imaging data acquired by the image sensor is fed back to the deformable mirror's surface control system. Through this feedback, the system can further adjust the actuation voltage to optimize the surface reconstruction accuracy.

[0076] Step S105: Based on the evaluation parameters, the surface reconstruction accuracy is improved by using a small step size and high-order search optimization algorithm so that the surface reconstruction accuracy reaches a theoretical optimal value of the search optimization algorithm.

[0077] It can be seen from the above steps that only when the residual error of the surface coefficient is less than the set threshold, the search optimization algorithm will be used to perform further search optimization through the feedback of the image sensor.

[0078] The surface coefficient residual is used as the first four-order surface characterization vector deviation e between the target surface and the current mirror surface n For example, when its e n When it is less than the threshold value t, further search optimization can be performed through a voltage search algorithm (such as a stochastic parallel gradient descent algorithm SPGD) to approach and stabilize at the theoretical optimal value. The entire control process of the above steps S101-S105 can be expressed by the following formula:

[0079]

[0080] In the above formula, e n The case when t is greater than or equal to t represents the processing in step S103, and e n The case when is less than t represents the processing in steps S104 and S105. Where E is the evaluation parameter of the SPGD search phase, and the energy gradient function is used here, and its expression is:

[0081]

[0082] In the above formula, E is the evaluation parameter, g(x,y) represents the grayscale value of the image at (x,y), and g(x+1,y) represents the grayscale value of the image at (x+1,y).

[0083] Preferably, when the above step S102 determines that the surface coefficient residual is less than the set threshold, the method of the present application may also include: after the system realizes the surface shape of the current focal length once, storing the key-value pair mapping relationship of "focal length-channel voltage" in a hash table to quickly realize system zoom.

[0084] After the system successfully achieves the current focal length surface shape adjustment, the key-value mapping relationship of "focal length-channel voltage" is stored in the hash table. This process ensures that when the same focal length needs to be reproduced in the future, the calculated channel voltage can be quickly read from the hash table and applied, thereby achieving fast zooming. Therefore, this method reduces the need to recalculate the voltage every time the zoom is changed, and improves the response speed and efficiency of the system.

[0085] Further preferably, when the above step S102 determines that the surface coefficient residual is less than the set threshold, the method of the present application may also include: after the system realizes the surface shape of the current focal length once, storing the key-value pair mapping relationship of "tilt angle-channel voltage" in a hash table to quickly realize system image stabilization.

[0086] Further preferably, after using a hash table to store the above-mentioned "focal length-channel voltage" and "tilt angle-channel voltage", and before using a wavefront sensor to monitor in real time the wavefront component after deformable mirror shape modulation, the method of the present application may also include: determining whether the target focal length or target tilt angle is in the hash table, if so, directly reading the corresponding channel voltage in the hash table, and controlling the deformable mirror based on the channel voltage; if not, using a wavefront sensor to monitor in real time the wavefront component after deformable mirror shape modulation.

[0087] This method reduces the need for real-time calculations by giving priority to pre-stored data in the hash table, thereby improving the efficiency and responsiveness of the system. At the same time, it retains the ability to monitor in real time to cope with focal lengths or tilt angles that are not stored, ensuring the flexibility and accuracy of the system.

[0088] Further preferably, the functions of the wavefront sensor and the image sensor in the above steps can be realized by using a high-precision interferometer. Because the high-precision interferometer can not only obtain the standard Zernike coefficient composition of the deformable mirror surface shape and play the role of a wavefront sensor, but also directly obtain the vector height information of the current surface shape as an evaluation parameter to replace the function of the image sensor. The double-feedback surface shape control method of the deformable mirror under the condition of using a high-precision interferometer can be expressed as Figure 3 As shown, the black solid line represents the hardware data link. At the beginning of the control, the interferometer feeds back the measured surface standard Zernike coefficients to the computer, and the computer performs voltage iteration based on the influence function matrix. When it is less than the conversion threshold, the interferometer feeds back the collected surface sag data to the computer, and uses the RMSE of the measured sag and the target sag as the evaluation parameter to start the voltage refinement search, thereby realizing the dual feedback surface control based on the high-precision interferometer deformable mirror.

[0089] The use of high-precision interferometers to replace the functions of wavefront sensors and image sensors can simplify the hardware configuration of the system. In addition, interferometers provide high-precision surface shape measurement capabilities and can accurately obtain subtle changes in surface shape. This high-precision measurement is essential for achieving accurate surface shape reconstruction and wavefront correction. By directly obtaining the sagittal height information of the surface shape, the interferometer can provide more accurate evaluation parameters for feedback control, which helps to improve the accuracy of surface shape reconstruction and make it close to the global optimum. Therefore, the use of high-precision interferometers not only simplifies the system structure, but also significantly improves the accuracy and efficiency of surface shape control, meeting the high requirements of the fixed zoom stabilization system.

[0090] The principle of the mirror convergence method based on dual feedback surface shape control optimization described in the above steps S101-S105 can also be found in Figure 4 As shown in the figure, the object light emitted (reflected) by the target at infinity can be approximated as parallel light. Due to the atmospheric turbulence and optical elements, a certain degree of wavefront distortion will be generated. The wavefront sensor (WFS) monitors the wavefront components after the deformable mirror surface shape modulation in real time, and feeds back the residual wavefront shape component coefficients to the deformable mirror. The image sensor (CMOS) detects the image of the target after wavefront distortion compensation in real time, and also feeds back the image quality parameters to the deformable mirror surface shape control, forming a dual-feedback surface shape control method.

[0091] The above method is further explained below by taking the fixed zoom image stabilization system as an example. Figure 5 FIG. 1 is a flow chart of a mirror convergence method in a fixed zoom image stabilization system provided by an embodiment of the present invention. Figure 5 It can be seen that after the zoom stabilization starts, the system first determines whether the target focal length or tilt angle is in the hash table: if it is, the voltage in the hash table is read, and the deformable mirror (DM) is executed to generate the corresponding deformation, and the image sensor is entered into the step of collecting and imaging; if it is not, the wavefront sensor detects the variable mirror shape in real time, and at the same time determines whether the 2 norm of the residual of the first four-order surface coefficient is less than the threshold. If it is less than the threshold, the image sensor is entered into the step of collecting and imaging, and the "focal length-channel voltage" and "tilt angle-channel voltage" are stored in the hash table. If it is not less than the threshold, the iterative voltage is calculated, and the DM is controlled to generate the corresponding deformation. After entering the step of collecting and imaging by the image sensor, it is determined whether the evaluation parameters are stable. If they are stable, the zoom stabilization ends. If they are unstable, a bidirectional voltage random disturbance is generated, which is executed by the DM respectively, and then the gradient is estimated and the control voltage is updated.

[0092] From the above, it can be seen that the mirror convergence method based on dual feedback surface shape control optimization provided by the present invention makes up for the defects that the control accuracy of the closed-loop control algorithm based on the influence function is affected by the measurement accuracy of the influence function matrix, and the closed-loop control algorithm without the influence function has low calculation efficiency and is easy to converge to the local optimal value. It can maintain the stability of control accuracy under the conditions of variable aperture and variable aberration surface shape and meet the requirements of the fixed zoom stabilization system.

[0093] The effect of the above method of the present application is further described below by taking a fixed zoom image stabilization system as an example through specific experimental data:

[0094] The fixed zoom stabilization system has changes in the aperture and target surface shape components at different focal lengths. In order to verify whether the existing surface shape control algorithm can maintain the stability of control accuracy under the condition of changing aperture, the pseudo-inverse method, SD, ILC, and SPGD simulations are used to control the deformable mirror to achieve the variable aperture first six-order Zernike aberration surface shape. At the same time, the root mean square error of the surface shape reconstruction result is recorded as the evaluation standard for the deformable mirror's fitting accuracy for the free-form surface. The smaller the root mean square error, the more accurate the mirror surface shape fitting. Taking 30% to 80% of the mirror aperture as the working aperture of the deformable mirror, and changing from small aperture to large aperture at a step frequency of 2%, the statistical results of the reconstruction error of the first six-order Zernike aberration surface shape under different apertures are shown as follows: Figure 6 As shown in the figure, the depth of the grayscale represents the root mean square error of the algorithm to reconstruct the target surface shape. The darker the color, the larger the root mean square error.

[0095] Depend on Figure 6 From the simulation results, it can be seen that the open-loop pseudo-inverse method has a large error in controlling the surface shape of small apertures and high orders, and is not suitable for high-precision application scenarios such as imaging systems; the overall reconstruction accuracy of the SD and ILC methods is relatively close, indicating that the two are essentially similar algorithms, and their reconstruction accuracy cannot be further improved due to the influence of the accuracy of the function matrix; SPGD has poor reconstruction accuracy for defocused (Z5) and spherical aberration (Z13) surface shapes, and the free-form surface that the deformable mirror needs to realize in the fixed zoom system is often dominated by the defocused surface shape, so SPGD is not suitable for use in zoom stabilization systems.

[0096] statistics Figure 6 The average values ​​of the reconstruction error to the Zernike terms and the target aperture respectively indicate the adaptability of the algorithm to the variable aberration surface shape and the variable optical aperture. The results are shown in Figure 7 shown.

[0097] Depend on Figure 7It can be seen that with the increase of the order of the target surface shape, the surface reconstruction error of the existing algorithms shows an upward trend. Among them, the open-loop pseudo-inverse method has the largest error when realizing complex surface shapes, and the SPGD method has a significantly improved error when realizing defocus and spherical aberration surface shapes. When the aperture changes, the solution errors of the pseudo-inverse method, SD and ILC show a trend of first decreasing, then increasing and then decreasing with the increase of the aperture. Among them, the pseudo-inverse method has a larger error when realizing the surface shape of a small aperture; the surface shape realization accuracy of SPGD is worse than the pseudo-inverse method when the target surface shape has a small aperture, but worse than SD and ILC. When the aperture is medium, the realization accuracy is the worst among the existing algorithms, and it is the best when the aperture is large. Comprehensive analysis shows that the above algorithms cannot meet the requirements of the fixed zoom stabilization system for the reconstruction accuracy of the target surface shape with variable aperture and variable aberration components.

[0098] Taking the computational cost of the open-loop pseudo-inverse method as the unit, the computational efficiency of the existing algorithms is compared, and the statistical results are as follows: Figure 8 As shown in Figure 2, it can be seen that the calculation efficiency of SD and ILC methods is higher, and the calculation efficiency of SPGD is the worst.

[0099] From the above analysis, it can be seen that the closed-loop control algorithm based on the influence function has good control efficiency, but its control accuracy is affected by the measurement accuracy of the influence function matrix, and the surface reconstruction accuracy cannot be further improved. The closed-loop control algorithm without the influence function can generally achieve a higher precision surface shape, but it often requires thousands of iterations to converge, and there is a possibility of converging to a local optimal value. The existing algorithms cannot maintain the stability of control accuracy under the condition of variable aperture and variable aberration surface shape, and cannot meet the requirements of the fixed zoom stabilization system.

[0100] After implementing the mirror convergence method based on dual feedback surface shape control optimization of this scheme, in order to explore the realization accuracy of the dual feedback surface shape control method when changing the optical aperture and the aberration surface shape, the root mean square error of the surface shape reconstruction is also recorded as follows: Fig. 9 As shown, DFSC is the mirror convergence method based on double feedback surface shape control optimization of the present application.

[0101] The statistical double feedback surface control method is used to calculate the average value of the root mean square error of the reconstruction of the surface with different aberration components at the same aperture, and the average value of the root mean square error of the reconstruction of the surface with the same aberration components at different apertures, and is compared with the existing methods. The statistical results are as follows Fig.10 shown.

[0102] It can be seen that compared with the existing algorithms, the control accuracy of the dual feedback surface control method proposed in this patent is optimal in the full range of aperture and the first six order aberration surface shapes, especially in the realization of the defocus surface shape, the reconstruction accuracy has been significantly improved, which proves the practicality of the algorithm used in this patent. It can adjust the optical focal length of the mirror with high precision and meet the requirements of constructing an offline control voltage hash set of a fixed zoom system.

[0103] Taking the calculation time of the pseudo-inverse method as the benchmark, the calculation cost of other existing algorithms and the method of this patent are standardized for comparison. The statistical results are as follows: Fig.11 The results show that the computational cost of the patented method is reduced by about 43% compared with SPGD, which is 6-8 times the computational cost of SD or ILC.

[0104] Realizing the defocused surface shapes with different actuation amounts is the basis of the system zoom, and realizing various free-form surface aberration shapes is the basis of the system image stabilization and image quality optimization. The dual feedback surface control method and iterative learning control method are used to simulate and reconstruct the defocused surface shapes with different actuation amounts and the different aberration surface shapes with the same actuation amounts, and the root mean square error is statistically analyzed. The results are as follows: Fig.12 shown.

[0105] The results show that with the increase of the actuation amount of the defocus item, the surface reconstruction errors of the two algorithms increase, but the error accumulation rate of ILC is faster. The DFCS method proposed in this patent can reduce the reconstruction root mean square error by more than 40%; the free-form surface shapes such as third and fifth order aberrations, coma, initial and high order spherical aberrations, trefoil, etc. common in optical systems are simulated and reconstructed. The actuation amount of the preset target surface shape is 8 microns. The simulation results show that the method of this patent can also significantly reduce the reconstruction root mean square error of the deformable mirror for the free-form surface by more than 50%, which is beneficial to the realization of free-form surfaces in zoom stabilization systems.

[0106] When the fixed zoom system uses a hash set to pre-store the "focal length-voltage" key-value pairs, the search time complexity is only O(1), which can be ignored. The zoom response delay is only related to the actuator response delay. The momentum deformable mirror actuator model used in the experimental verification of the surface shape control optimization algorithm of this patent is PPA80L, and the capacitor C piezo =5.4μF, the driving power supply is CILAS's LA75B, and its limit output current is I lim =360mA. When the actuator is working dynamically, the current and signal frequency are approximately linearly related, as shown in the following formula:

[0107] i≈2πfC piezo V P

[0108] Where V P is the middle value of the control voltage range. For PPA80L, V P =85V. Therefore, the maximum operating frequency of the system can be expressed as follows:

[0109]

[0110] Will I lim 、V P , C piezoSubstituting the value of into the above formula, the maximum operating frequency is about 125Hz. Therefore, the actuator takes time t = 1 / f to achieve full stroke actuation. max =8ms, which meets the requirement of the fixed zoom system for ms-level zoom.

[0111] It can be clearly seen from the above experimental data that the mirror convergence method based on dual feedback surface shape control optimization of the present application makes up for the fact that the control accuracy of the closed-loop control algorithm based on the influence function is affected by the measurement accuracy of the influence function matrix, and the closed-loop control algorithm without the influence function has low computational efficiency and is easy to converge to the local optimal value. It can maintain the stability of control accuracy under the condition of variable aperture and variable aberration surface shape and meet the requirements of the fixed zoom stabilization system.

[0112] like Fig.13 The structure diagram of a mirror convergence device based on dual feedback surface shape control optimization provided by an embodiment of the present application is shown, and the device includes: a wavefront monitoring unit 130, a judgment unit 131, a fast reconstruction unit 132, an image monitoring unit 133 and a secondary reconstruction unit 134, wherein the judgment unit 131 is connected to the wavefront monitoring unit 130, the fast reconstruction unit 132 and the image monitoring unit 133 respectively, and the secondary reconstruction unit 134 is connected to the image monitoring unit 133. Among them:

[0113] The wavefront monitoring unit 130 is used to monitor the wavefront components modulated by the deformable mirror in real time using a wavefront sensor.

[0114] The judging unit 131 is used to judge whether the surface coefficient residual of the wavefront component is greater than or equal to a set threshold.

[0115] The fast reconstruction unit 132 is used to quickly obtain the surface reconstruction voltage through the influence function matrix when the judgment unit judges that the surface coefficient residual is greater than or equal to the set threshold, and make the deformable mirror converge quickly based on the surface reconstruction voltage.

[0116] The image monitoring unit 133 is used to use an image sensor to monitor changes in light intensity distribution or imaging clarity as evaluation parameters for the degree of distortion wavefront correction when the judgment unit judges that the surface coefficient residual is less than the set threshold.

[0117] The secondary reconstruction unit 134 is used to improve the surface reconstruction accuracy based on the evaluation parameters through a small step size and high-order search optimization algorithm so that the surface reconstruction accuracy reaches the theoretical optimal value of the search optimization algorithm.

[0118] Preferably, Fig.14As shown, the above-mentioned device also includes: a storage unit 135, which is used to store the key-value pair mapping relationship of "focal length-channel voltage" in a hash table when the judgment unit judges that the surface coefficient residual is less than the set threshold value, and after the system has realized the surface shape of the current focal length once, so as to quickly realize system zoom.

[0119] Preferably, the above-mentioned storage unit 135 is also used to store the key-value pair mapping relationship of "tilt angle-channel voltage" in a hash table when the judgment unit judges that the surface coefficient residual is less than the set threshold value, and after the system has realized the surface shape of the current focal length once, so as to quickly realize system image stabilization.

[0120] Preferably, Fig.15 As shown, the above-mentioned device also includes: a hash table judgment unit 136, which is used to judge whether the target focal length or target inclination angle is in the hash table. If so, the corresponding channel voltage in the hash table is directly read, and the deformable mirror is controlled based on the channel voltage; if not, the wavefront sensor is used to monitor the wavefront component after the deformable mirror surface shape is modulated in real time.

[0121] Preferably, the fast reconstruction unit 132 quickly obtains the surface reconstruction voltage through the influence function matrix, including: using the surface characterization vector difference between the target surface and the current surface as an evaluation function, and the evaluation function is shown in the following formula:

[0122] J(x)=e n =S d -S n =S d -GC n

[0123] In the above formula, J(x) is the evaluation function, e n is the first four-order surface characterization vector deviation between the target surface and the current surface, S d is the target surface characterization vector, C n is the control voltage signal currently acting on the deformable mirror, S n is the characterization vector of the current deformed mirror shape, and G is the influence function matrix.

[0124] Under the guidance of the gradient descent method, the reconstructed voltage signal is adjusted in the opposite direction of the gradient of the evaluation function, and the reconstructed voltage signal of the deformable mirror in the next iteration is:

[0125] C n+1 =C n +αG T e n

[0126] In the above formula, C n+1 Reconstruct the voltage signal of the deformable mirror for the next iteration, and α is the gain coefficient.

[0127] Preferably, the above search optimization algorithm is a stochastic parallel gradient descent algorithm, and the evaluation parameters are:

[0128]

[0129] In the above formula, E is the evaluation parameter, g(x,y) represents the grayscale value of the image at (x,y), and g(x+1,y) represents the grayscale value of the image at (x+1,y).

[0130] Preferably, the wavefront sensor and the image sensor in the above device are implemented using a high-precision interferometer.

[0131] The detailed description of each of the above units can be found in the corresponding description in the aforementioned method embodiment, which will not be further elaborated here.

[0132] It can be seen from the above technical scheme that the mirror convergence device based on dual feedback surface shape control optimization provided by the present invention makes up for the defects that the control accuracy of the closed-loop control algorithm based on the influence function is affected by the measurement accuracy of the influence function matrix, and the closed-loop control algorithm without the influence function has low calculation efficiency and is easy to converge to the local optimal value. It can maintain the stability of control accuracy under the condition of variable aperture and variable aberration surface shape, and meet the requirements of the fixed zoom stabilization system.

[0133] An embodiment of the present invention further provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the above method is implemented when the processor executes the program.

[0134] An embodiment of the present invention further provides a computer-readable storage medium, which stores a computer program for executing the above method.

[0135] like Fig.16 As shown, the electronic device 600 may further include: a communication module 110, an input unit 120, an audio processor 130, a display 160, and a power supply 170. It is worth noting that the electronic device 600 does not necessarily have to include Fig.16 In addition, the electronic device 600 may also include Fig.16 For components not shown, reference may be made to the prior art.

[0136] like Fig.16 As shown, the central processor 100 is sometimes also referred to as a controller or an operation control, and may include a microprocessor or other processor devices and / or logic devices. The central processor 100 receives inputs and controls the operations of various components of the electronic device 600.

[0137] The memory 140 may be, for example, one or more of a cache, a flash memory, a hard drive, a removable medium, a volatile memory, a non-volatile memory or other suitable devices. The above-mentioned information related to the failure may be stored, and a program for executing the relevant information may also be stored. The CPU 100 may execute the program stored in the memory 140 to implement information storage or processing.

[0138] The input unit 120 provides input to the CPU 100. The input unit 120 is, for example, a key or a touch input device. The power supply 170 is used to provide power to the electronic device 600. The display 160 is used to display display objects such as images and text. The display may be, for example, an LCD display, but is not limited thereto.

[0139] The memory 140 may be a solid-state memory, such as a read-only memory (ROM), a random access memory (RAM), a SIM card, etc. It may also be a memory that saves information even when the power is off, can be selectively erased, and is provided with more data, examples of which are sometimes referred to as EPROMs, etc. The memory 140 may also be some other type of device. The memory 140 includes a buffer memory 141 (sometimes referred to as a buffer). The memory 140 may include an application / function storage unit 142, which is used to store application programs and function programs or processes for executing the operation of the electronic device 600 through the central processor 100.

[0140] The memory 140 may also include a data storage unit 143 for storing data, such as contacts, digital data, pictures, sounds, and / or any other data used by the electronic device. The driver storage unit 144 of the memory 140 may include various drivers for communication functions of the electronic device and / or for executing other functions of the electronic device (such as messaging applications, address book applications, etc.).

[0141] The communication module 110 is a transmitter / receiver that transmits and receives signals via the antenna 111. The communication module 110 (transmitter / receiver) is coupled to the central processor 100 to provide input signals and receive output signals, which may be the same as the case of a conventional mobile communication terminal.

[0142] Based on different communication technologies, multiple communication modules 110 may be provided in the same electronic device, such as a cellular network module, a Bluetooth module and / or a wireless local area network module. The communication module 110 (transmitter / receiver) is also coupled to a speaker 1301 and a microphone 1302 via an audio processor 1300 to provide an audio output via the speaker 1301 and receive an audio input from the microphone 1302, thereby realizing a common telecommunication function. The audio processor 1300 may include any suitable buffer, decoder, amplifier, etc. In addition, the audio processor 1300 is also coupled to the central processor 100, so that recording can be performed on the local machine through the microphone 1302, and the sound stored on the local machine can be played through the speaker 1301.

[0143] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0144] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0145] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0146] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0147] The present invention uses specific embodiments to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A mirror convergence method based on double feedback surface shape control optimization, characterized in that: The method comprises: The wavefront sensor is used to monitor the wavefront components modulated by the deformable mirror in real time; Determining whether the surface coefficient residual of the wavefront component is greater than or equal to a set threshold; If the residual of the surface shape coefficient is greater than or equal to the set threshold, the surface shape reconstruction voltage is quickly obtained through the influence function matrix, and the deformable mirror is quickly converged based on the surface shape reconstruction voltage, and then the process returns to continue to determine whether the residual of the surface shape coefficient of the wavefront component is greater than or equal to the set threshold; If the surface coefficient residual is less than the set threshold, the image sensor is used to monitor the change of light intensity distribution or imaging clarity as an evaluation parameter of the degree of distortion wavefront correction; Based on the evaluation parameters, the surface reconstruction accuracy is improved by using a small step size and high number of search optimization algorithms so that the surface reconstruction accuracy reaches a theoretical optimal value of the search optimization algorithm.

2. The mirror convergence method based on dual feedback surface shape control optimization according to claim 1, characterized in that: If the surface coefficient residual is less than the set threshold, the method further includes: after the system realizes the surface shape of the current focal length once, storing the key-value pair mapping relationship of "focal length-channel voltage" in a hash table to quickly realize system zoom.

3. The mirror convergence method based on dual feedback surface shape control optimization according to claim 2, characterized in that: If the surface coefficient residual is less than the set threshold, the method further includes: after the system realizes the surface shape of the current focal length once, storing the key-value pair mapping relationship of "tilt angle-channel voltage" in a hash table to quickly realize system image stabilization.

4. The mirror convergence method based on dual feedback surface shape control optimization according to claim 3, characterized in that: Before using the wavefront sensor to monitor the wavefront component modulated by the deformable mirror in real time, the method further includes: It is determined whether the target focal length or target inclination angle is in the hash table. If so, the corresponding channel voltage in the hash table is directly read, and the deformable mirror is controlled based on the channel voltage; if not, the wavefront sensor is used to monitor the wavefront component after the deformable mirror surface shape is modulated in real time.

5. The mirror convergence method based on dual feedback surface shape control optimization according to claim 1, characterized in that: The method of quickly obtaining the surface reconstruction voltage by using the influence function matrix comprises: The difference between the target face shape and the current face shape is used as the evaluation function, and the evaluation function is shown in the following formula: J(x)=e n =S d -S n =S d -GC n In the above formula, J(x) is the evaluation function, e n is the first four-order surface characterization vector deviation between the target surface and the current surface, S d is the target surface characterization vector, C n is the control voltage signal currently acting on the deformable mirror, S n is the characterization vector of the current deformed mirror shape, and G is the influence function matrix; Under the guidance of the gradient descent method, the reconstructed voltage signal is adjusted in the opposite direction of the gradient of the evaluation function, and the reconstructed voltage signal of the deformable mirror in the next iteration is: C n+1 =C n +αG T e n In the above formula, C n+1 Reconstruct the voltage signal of the deformable mirror for the next iteration, and α is the gain coefficient.

6. The mirror convergence method based on dual feedback surface shape control optimization according to claim 1, characterized in that: The search optimization algorithm is a stochastic parallel gradient descent algorithm, and the evaluation parameters are: E=∑ y ∑ x (g(x+1,y)-g(x,y) 2 +g(x,y+1)-g(x,y) 2 ) In the above formula, E is the evaluation parameter, g(x,y) represents the grayscale value of the image at (x,y), and g(x+1,y) represents the grayscale value of the image at (x+1,y).

7. The mirror convergence method based on dual feedback surface shape control optimization according to claim 1, characterized in that: The wavefront sensor and the image sensor are implemented using a high-precision interferometer.

8. A mirror convergence device based on dual feedback surface shape control optimization, characterized in that: The device comprises: A wavefront monitoring unit, used to monitor the wavefront components modulated by the deformable mirror in real time using a wavefront sensor; A judging unit, used to judge whether the residual of the surface coefficient of the wavefront component is greater than or equal to a set threshold; A fast reconstruction unit, configured to quickly obtain a surface reconstruction voltage through an influence function matrix when the judgment unit judges that the surface coefficient residual is greater than or equal to the set threshold, and to make the deformable mirror converge quickly based on the surface reconstruction voltage; An image monitoring unit, configured to use an image sensor to monitor changes in light intensity distribution or imaging clarity as evaluation parameters for the degree of distortion wavefront correction when the judging unit judges that the surface coefficient residual is less than the set threshold value; The secondary reconstruction unit is used to improve the surface reconstruction accuracy based on the evaluation parameters through a small step size and high-order search optimization algorithm so that the surface reconstruction accuracy reaches the theoretical optimal value of the search optimization algorithm.

9. The mirror convergence device based on dual feedback surface shape control optimization according to claim 8, characterized in that: The device also includes: a storage unit, which is used to store the key-value pair mapping relationship of "focal length-channel voltage" in a hash table when the judgment unit judges that the surface coefficient residual is less than the set threshold and after the system has realized the surface shape of the current focal length once, so as to quickly realize system zoom.

10. The mirror convergence device based on dual feedback surface shape control optimization according to claim 9, characterized in that: The storage unit is also used to store the key-value pair mapping relationship of "tilt angle-channel voltage" in a hash table when the judgment unit judges that the surface coefficient residual is less than the set threshold and after the system has realized the surface shape of the current focal length once, so as to quickly realize system stabilization.

11. The mirror convergence device based on dual feedback surface shape control optimization according to claim 10, characterized in that: The device also includes: The hash table judgment unit is used to judge whether the target focal length or target inclination angle is in the hash table. If so, the corresponding channel voltage in the hash table is directly read, and the deformable mirror is controlled based on the channel voltage; if not, the wavefront sensor is used to monitor the wavefront component after the deformable mirror surface shape is modulated in real time.

12. The mirror convergence device based on dual feedback surface shape control optimization according to claim 8, characterized in that: The fast reconstruction unit quickly obtains the surface reconstruction voltage through the influence function matrix, including: The difference between the target face shape and the current face shape is used as the evaluation function, and the evaluation function is shown in the following formula: J(x)=e n =S d -S n =S d -GC n In the above formula, J(x) is the evaluation function, e n is the first four-order surface characterization vector deviation between the target surface and the current surface, S d is the target surface characterization vector, C n is the control voltage signal currently acting on the deformable mirror, S n is the characterization vector of the current deformed mirror shape, and G is the influence function matrix; Under the guidance of the gradient descent method, the reconstructed voltage signal is adjusted in the opposite direction of the gradient of the evaluation function, and the reconstructed voltage signal of the deformable mirror in the next iteration is: C n+1 =C n +αG T e n In the above formula, C n+1 Reconstruct the voltage signal of the deformable mirror for the next iteration, and α is the gain coefficient.

13. The mirror convergence device based on dual feedback surface shape control optimization according to claim 8, characterized in that: The search optimization algorithm is a stochastic parallel gradient descent algorithm, and the evaluation parameters are: E=∑ y ∑ x (g(x+1,y)-g(x,y) 2 +g(x,y+1)-g(x,y) 2 ) In the above formula, E is the evaluation parameter, g(x,y) represents the grayscale value of the image at (x,y), and g(x+1,y) represents the grayscale value of the image at (x+1,y).

14. The mirror convergence device based on dual feedback surface shape control optimization according to claim 8, characterized in that: The wavefront sensor and the image sensor are implemented using a high-precision interferometer.

15. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

16. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images