A wavefront aberration reconstruction method, device, electronic device and storage medium

Abstract

The present invention discloses a wave aberration reconstruction method, apparatus, electronic device, and storage medium, relating to the technical field of optical detection. The wave aberration reconstruction method includes: obtaining a corresponding initial Zernike coefficient set based on a pre-acquired actual point spread function image and a first training network; optimizing the initial Zernike coefficient set by using a second training network based on the principle that the target point spread function image is the same as the actual point spread function image to obtain a corresponding target Zernike coefficient set; and obtaining a corresponding target wave aberration based on the target Zernike coefficient set and a pre-configured optical system basis function. The present invention solves the problems that wave aberration reconstruction requires multiple measurements and calibrations, resulting in inconvenient operation and low accuracy in the prior art, thereby simply and conveniently completing the process of wave aberration reconstruction, saving manpower and time, and improving the reconstruction accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of optical detection, and in particular, to a wave aberration reconstruction method, apparatus, electronic device, and storage medium. Background Art

[0002] Wave aberration refers to the deviation between the actual wavefront and the ideal wavefront, which determines the size of the point spread function of the system, and further affects the aberration and imaging quality of the system. Characterizing and correcting wave aberration is crucial in many fields, such as adaptive optics, astronomical telescopes, microscopes, and optical communication.

[0003] Currently, the existing wave aberration reconstruction methods mainly include two types: fringe deflection technique and deep learning-based methods. The method based on fringe deflection technique requires measuring the deformation of fringe projections with and without the test object; the deep learning-based method is usually a supervised method, and it is difficult to obtain high-quality training data because measuring the accurate point spread function and wave aberration at multiple fields of view and wavelengths is time-consuming and laborious.

[0004] However, the existing wave aberration reconstruction methods have some deficiencies. The method based on fringe deflection technique usually requires multiple measurements and precise system alignment and calibration, which is inconvenient to operate; the deep learning-based method has low accuracy. Summary of the Invention

[0005] The present invention provides a wave aberration reconstruction method, apparatus, electronic device, and storage medium, which solves the problems of inconvenient operation caused by multiple measurements and calibration for wave aberration reconstruction and low accuracy.

[0006] In a first aspect, an embodiment of the present invention provides a wave aberration reconstruction method, including:

[0007] Based on a pre-acquired actual point spread function image and a first training network, obtaining a corresponding initial set of Zernike coefficients;

[0008] Based on the principle that the target point spread function image is the same as the actual point spread function image, and using a second training network to optimize the initial set of Zernike coefficients, obtaining a corresponding target set of Zernike coefficients;

[0009] Based on the target set of Zernike coefficients and a pre-configured optical system basis function, obtaining a corresponding target wave aberration.

[0010] In a second aspect, an embodiment of the present invention further provides a wave aberration reconstruction apparatus, including:

[0011] An initial set acquisition module, configured to obtain a corresponding initial set of Zernike coefficients based on a pre-acquired actual point spread function image and a first training network;

[0012] A target set obtaining module, configured to optimize an initial set of Zernike coefficients based on the principle that the target point spread function image is the same as the actual point spread function image, and use a second training network to obtain a corresponding target set of Zernike coefficients;

[0013] A target wave aberration obtaining module, configured to obtain a corresponding target wave aberration based on the target set of Zernike coefficients and a pre-configured optical system basis function.

[0014] In a third aspect, an embodiment of the present invention further provides an electronic device, including:

[0015] At least one processor; and

[0016] A memory communicatively connected to the at least one processor; wherein,

[0017] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the wave aberration reconstruction method according to any one of the embodiments of the present invention.

[0018] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions for causing a processor to implement the wave aberration reconstruction method according to any one of the embodiments of the present invention when executed.

[0019] Based on the principle that the target point spread function image is the same as the actual point spread function image, and using the best imaging plane corresponding to the pre-configured optical system basis function to measure the corresponding target point spread function image, this embodiment solves the problem of inconvenient operation caused by multiple measurements and calibrations required for wave aberration reconstruction, thus simply and conveniently completing the process of wave aberration reconstruction, saving manpower and time, and improving the efficiency of wave aberration reconstruction; moreover, two training networks are used for wave aberration reconstruction, solving the problem of low accuracy in the prior art and improving the reconstruction accuracy.

[0020] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present invention, nor is it used to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other drawings based on these drawings without creative efforts.

[0022] Figure 1Flowchart of a wavefront aberration reconstruction method provided by an embodiment of the present invention;

[0023] Figure 2 Schematic diagram of a method for synthesizing a dataset provided by an embodiment of the present invention;

[0024] Figure 3 Visualization schematic diagram of a PSF solution space provided by an embodiment of the present invention;

[0025] Figure 4 Flowchart of another wavefront aberration reconstruction method provided by an embodiment of the present invention;

[0026] Figure 5 Schematic diagram of the structure of a second training network provided by an embodiment of the present invention;

[0027] Figure 6 Another implementation schematic of wavefront aberration reconstruction provided by an embodiment of the present invention;

[0028] Figure 7 Comparison schematic diagram of wavefront aberration reconstruction results obtained by using different implementation schemes provided by an embodiment of the present invention;

[0029] Figure 8 Schematic diagram of the structure of a wavefront aberration reconstruction device provided according to an embodiment of the present invention;

[0030] Figure 9 Schematic diagram of the structure of an electronic device for implementing the wavefront aberration reconstruction method of an embodiment of the present invention. Detailed implementation manners

[0031] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts shall fall within the protection scope of the present invention.

[0032] It should be noted that the terms "first", "second", etc. in the description, claims and the above drawings of the present invention are used to distinguish similar objects, and do not necessarily have to be used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0033] In one embodiment, Figure 1 is a flowchart of a wavefront aberration reconstruction method provided by an embodiment of the present invention. This embodiment is applicable to the situation of reconstructing wavefront aberration. This method can be executed by a wavefront aberration reconstruction device, and the wavefront aberration reconstruction device can be implemented in the form of hardware and / or software.

[0034] As Figure 1 shown, a wavefront aberration reconstruction method provided by this embodiment may include:

[0035] S110. Based on the pre-acquired actual point spread function image and the first training network, obtain the corresponding initial set of Zernike coefficients.

[0036] In the embodiment of the present invention, the actual point spread function image (Point Spread Function, PSF) refers to the light field distribution image of the output image when the input object is a point light source. Exemplarily, the actual point spread function image can be the spot measured at the best imaging plane. Figure 2 is a schematic diagram of a method for synthesizing a data set provided by an embodiment of the present invention. As Figure 2 shown, the actual point spread function image can be understood as a synthetic data set of the point spread function and the wavefront aberration data pair. Figure 3 is a visualization schematic diagram of a PSF solution space provided by an embodiment of the present invention, where other Zernike coefficients are 0. As Figure 3 shown, the PSF solution space is non-convex. Due to the rotational symmetry of the PSF, only the PSF falling on the +y axis can be selected. The PSF and the wavefront aberration at this time can be taken out as a pair of data pairs.

[0037] The first training network can be understood as a convolutional neural network for wavefront aberration prediction, and the first training network can be a supervised learning network. For example, the first training network can be a U-Net neural network. The initial Zernike coefficient set refers to a set of Zernike coefficients directly obtained based on the actual point spread function image and the prediction of the first training network. The initial Zernike coefficient set can be used to map the wavefront aberration corresponding to the actual point spread function, or project the wavefront aberration corresponding to the actual point spread function into the Zernike coefficient space to obtain the initial Zernike coefficient set. Specifically, the pre-obtained actual point spread function image can be input into a suitable learning network for training to obtain the initial Zernike coefficient set corresponding to the actual point spread function image.

[0038] In one embodiment, obtaining the corresponding initial Zernike coefficient set based on the pre-obtained actual point spread function image and the first training network may include:

[0039] Input the actual point spread function image into the first training network to obtain the corresponding initial wavefront aberration;

[0040] Project the initial wavefront aberration into the Zernike coefficient space to obtain the corresponding initial Zernike coefficient set.

[0041] Specifically, the actual point spread function image can be input into the first training network for training and prediction of the first training network, and the initial wavefront aberration corresponding to the actual point spread function image is output. The initial wavefront aberration of the actual point spread function image can be projected into the Zernike coefficient space for calculation to obtain a set of Zernike coefficients, and this set of Zernike coefficients can be used as the initial Zernike coefficient set corresponding to the actual point spread function image.

[0042] S120. Based on the principle that the target point spread function image is the same as the actual point spread function image, and using the second training network to optimize the initial Zernike coefficient set to obtain the corresponding target Zernike coefficient set.

[0043] In the embodiment of the present invention, the second training network is used to optimize the initial Zernike coefficient set and obtain the corresponding target Zernike coefficient set by minimizing the self-supervised loss. Exemplarily, the second training network can be a self-supervised learning network; the target point spread function image can be understood as the final point spread function image output after training by the second training network, and this target point spread function image is the same as the actual point spread function image or the loss value between the two is the smallest. The target Zernike coefficient set refers to a set of Zernike coefficients optimized by the second training network.

[0044] Specifically, based on the principle that the finally output point spread function image from training is the same as the actual point spread function image, the second training network can be used to optimize the initial set of Zernike coefficients. The initial set of Zernike coefficients can be optimized by minimizing the difference between the target point spread function and the actual point spread function image, and the corresponding target set of Zernike coefficients can be obtained. Exemplarily, the second training network is calculated as follows:

[0045]

[0046] where, θ2 is a second training network before optimization; is the second training network after optimization; PSF is the actual point spread function image, which can also be understood as a reference point spread function image;

[0047] is a set of Zernike coefficients; K is used to represent the number of Zernike coefficients included in a set of Zernike coefficients (i.e., a set of Zernike coefficients). For example, due to the rotational symmetry of the PSF space, it is specified that the PSF falls on the +Y axis, that is, the other 16 terms are 0, and, at most, the number of Zernike coefficients can be 37. Correspondingly, K can be 21; is the target point spread function image. The process of continuously iteratively training the second training network can be understood as the process in which the actual loss value between the target point spread function image converted from the target set of Zernike coefficients output by the second training network and the actual point spread function image reaches the minimum.

[0048] In one embodiment, the principle that the target point spread function image is the same as the actual point spread function image may include:

[0049] Determine the actual loss value between the target point spread function image and the actual point spread function image;

[0050] When the actual loss value reaches the minimum, determine that the target point spread function image is the same as the actual point spread function image.

[0051] In the embodiments of the present invention, the actual loss value can be understood as the difference between the target point spread function image and the actual point spread function image.

[0052] Specifically, the difference between the target point spread function image obtained from the target wave aberration and the actual point spread function image can be determined as the actual loss value, and the actual loss value is calculated as follows:

[0053] Continuously train the second training network with this actual loss value, and compare the obtained new target point spread function image with the actual point spread function image. When the actual loss value reaches the minimum, it can be determined that the target point spread function image is the same as the actual point spread function image. The calculation formula for the minimum value of the actual loss value is as follows:

[0054]

[0055] Wherein, is a set of Zernike coefficients, K = 21, representing the number of Zernike coefficients, is the target point spread function image, and PSF is the actual point spread function image, represents the optimal set of target Zernike coefficients. That is to say, through the second training network, the optimal target Zernike coefficients can be output, and the loss value between the target point spread function image obtained by converting through the set of target Zernike coefficients and the actual point spread function image reaches the minimum. In the actual operation process, the set of target Zernike coefficients can include one Zernike coefficient or multiple Zernike coefficients, but the second training network will only output a set of Zernike coefficients.

[0056] S130. Obtain the corresponding target wave aberration based on the set of target Zernike coefficients and the pre-configured optical system basis function.

[0057] In the embodiment of the present invention, the target wave aberration refers to the wave aberration obtained through the second training network, and this wave aberration corresponds to the wave aberration of the actual point spread function image. The optical system basis function refers to a pre-configured basis function that can reconstruct the wavefront or wave aberration. For example, the optical system basis function can include but is not limited to one of the following: Zernike basis, Seidel basis. Generally speaking, the optical system basis function can have prior information and can represent the wave aberration with a small number of parameters, thereby avoiding excessive computational complexity.

[0058] Specifically, the target wave aberration corresponding to the target point spread function image can be calculated according to the obtained set of target Zernike coefficients above. The calculation method can include the weighted linear combination of the set of target Zernike coefficients and the optical system basis function. Exemplarily, based on the obtained set of target Zernike coefficients and the pre-configured optical system basis function, a weighted linear combination can be performed to calculate the corresponding target wave aberration, and the target point spread function image corresponding to this target wave aberration is the same as the actual point spread function image.

[0059] Based on the principle that the target point spread function image is the same as the actual point spread function image, this embodiment measures the corresponding target point spread function image using the best imaging plane corresponding to the pre-configured optical system basis function, thereby solving the problem of inconvenient operation caused by the need for multiple measurements and calibrations in wavefront aberration reconstruction, and thus simply and conveniently completing the process of wavefront aberration reconstruction, saving manpower and time, and improving the efficiency of wavefront aberration reconstruction; moreover, two training networks are used for wavefront aberration reconstruction, solving the problem of low accuracy in the prior art and improving the reconstruction accuracy.

[0060] In one embodiment, Figure 4 It is a flowchart of another wavefront aberration reconstruction method provided by an embodiment of the present invention. Based on the above embodiments, this embodiment further optimizes and expands the process of obtaining the target Zernike coefficient set and the target wavefront aberration of the target point spread function image based on the actual point spread function image, the first training network, and the second training network. The process of iteratively training the second training network can also be understood as a process of continuously adjusting the weight coefficients in the second training network.

[0061] As Figure 4 shown, another wavefront aberration reconstruction method provided by this embodiment may include:

[0062] S210. Based on the pre-acquired actual point spread function image and the first training network, obtain the corresponding initial Zernike coefficient set.

[0063] S220. Adjust the initial weight coefficients of the second training network based on the initial Zernike coefficient set to obtain the corresponding intermediate weight coefficients.

[0064] In the embodiment of the present invention, the intermediate weight coefficients can be understood as the network structure parameters of the second training network obtained by training using the initial Zernike coefficient set.

[0065] Specifically, the second training network can be trained based on the initial Zernike coefficient set input to the second training network. During this process, the initial weight coefficients of the second training network are continuously adjusted to obtain the intermediate weight coefficients of the second training network based on the initial Zernike coefficient set.

[0066] S230. Based on the second training network and the intermediate weight coefficients, obtain the corresponding intermediate Zernike coefficient set.

[0067] In the embodiment of the present invention, the intermediate Zernike coefficient set can be understood as a set of Zernike coefficients output by the second training network after the initial Zernike coefficient set is input to the second training network for initial weight coefficient adjustment.

[0068] Specifically, the adjusted intermediate weight coefficients and the initial Zernike coefficient set can be input into the second training network for training to obtain the intermediate Zernike coefficient set corresponding to the initial Zernike coefficient set.

[0069] S240. Obtain the corresponding target point spread function image based on the intermediate Zernike coefficient set.

[0070] Specifically, the corresponding target point spread function image can be calculated based on the obtained intermediate Zernike coefficient set above.

[0071] In one embodiment, S240 includes S2401 - S2403:

[0072] S2401. Perform a weighted linear combination based on the intermediate Zernike coefficient set and the pre - configured optical system basis function to obtain the corresponding intermediate wave aberration.

[0073] Specifically, a weighted linear combination can be performed based on the intermediate Zernike coefficient set and the pre - configured optical system basis function to obtain the corresponding intermediate wave aberration. Exemplarily, a fixed two - dimensional distribution can be used, multiplied and added with a weight to obtain the intermediate wave aberration. The obtained intermediate wave aberration is smooth, with finite peaks and valleys, and no such distributions as noise. The calculation of the intermediate wave aberration is as follows:

[0074]

[0075] Where Z is a fixed two - dimensional distribution. For example, Z can be a plane, hat - shaped or wave - shaped; n is the radial order, m is the azimuthal frequency, |m| ≤ n and n - |m| must be even, ρ is the radius coordinate of a point in the pupil area, and θ is the azimuth angle of the pupil plane; is the Zernike polynomial, and different orders of the Zernike polynomial correspond to different types of aberrations of the imaging system.

[0076] S2402. Determine the corresponding generalized pupil function based on the intermediate wave aberration, the pre - configured aperture function, and the wavelength.

[0077] Specifically, the corresponding generalized pupil function can be determined based on the obtained intermediate wave aberration, the pre - configured aperture function, and the wavelength. The calculation of the generalized pupil function is as follows:

[0078]

[0079] Where A(x, y) is the aperture function, W(x, y) is the wave aberration, and λ is the wavelength.

[0080] S2403. Use the square of the magnitude of the Fourier transform of the generalized pupil function as the corresponding target point spread function image.

[0081] Specifically, based on the square of the magnitude of the Fourier transform of the generalized pupil function, as the corresponding target point spread function image, the target point spread function image is calculated as follows:

[0082]

[0083] Wherein, F(P(x,y)) represents the result of performing a Fourier transform on the generalized pupil function P(x,y), and the target point spread function image PSF is obtained based on the square of the absolute value of the Fourier transform of the generalized pupil function P(x,y).

[0084] S250. Based on the principle that the target point spread function image is the same as the actual point spread function image, adjust the intermediate weight coefficients of the second training network to obtain the corresponding target weight coefficients.

[0085] In the embodiments of the present invention, the target weight coefficients can be understood as the network parameters of the second training network after adjustment. The target point spread function image obtained by training the second training network with the target weight coefficients is the same as the actual point spread function image.

[0086] Specifically, based on the principle that the target point spread function image is the same as the actual point spread function image, continuously adjust the intermediate weight coefficients of the second training network until the target point spread function image output by the second training network is the same as the actual point spread function image, and obtain the network parameters of the second training network at this time as the target weight coefficients.

[0087] S260. Obtain the corresponding target Zernike coefficient set based on the second training network and the target weight coefficients. Specifically, Figure 5 is a schematic structural diagram of a second training network provided by the embodiments of the present invention. The second training network incorporates a statistical constraint on the Zernike coefficients. Different orders of Zernike polynomials correspond to different types of optical aberrations. The target Zernike coefficient set can be obtained through this network, as Figure 5 shown. A fixed value can be input through 16 channels (16channelinput) and 11 channels (11channel input). The linear layer (Linear Layer) of the second training network is activated using an activation function (Leaky Relu), and linear calculations, non-linear tanh functions, and tanh functions with a failure probability (tanh with a failure probability) calculations and multiplication calculations (Multiply) are performed to obtain the corresponding target Zernike coefficient set.

[0088] S270. Obtain the corresponding target wave aberration based on the target Zernike coefficient set and the pre-configured optical system basis function.

[0089] According to the technical solution of the embodiment of the present invention, an initial Zernike coefficient set corresponding to the actual point spread function image is obtained through the first training network. Based on the initial Zernike coefficient set, the initial weight coefficients of the second training network are adjusted to obtain the corresponding intermediate weight coefficients and intermediate Zernike coefficient set, as well as the target point spread function image corresponding to the intermediate Zernike coefficient set. Based on the principle that the target point spread function image is the same as the actual point spread function image, the intermediate weight coefficients of the second training network are adjusted to obtain the corresponding target weight coefficients and target Zernike coefficient set. Based on the target Zernike coefficient set and the pre-configured optical system basis function, the corresponding target wave aberration is obtained. Thus, the problem that wave aberration reconstruction requires multiple measurements and calibrations, resulting in inconvenient operation, is solved. Therefore, the process of wave aberration reconstruction is completed simply and conveniently, saving manpower and time, and improving the efficiency of wave aberration reconstruction. Moreover, two training networks are used for wave aberration reconstruction, solving the problem of low accuracy in the prior art and improving the reconstruction accuracy.

[0090] In one embodiment, Figure 6 FIG. is another implementation schematic diagram of wave aberration reconstruction provided by the embodiment of the present invention. The method consists of a two-stage network. The first training network is a learning network. By inputting the actual point spread function image into the first training network, the corresponding predicted initial wave aberration can be obtained. The second training network further optimizes the initial Zernike coefficients through an optimization method. Exemplarily, the actual point spread function image is denoted as input PSF, the first training network is denoted as LWNet Stage I, the initial wave aberration is denoted as W0(x,y), the initial Zernike coefficient set is denoted as InitializedZernike Coeff icients, the second training network is denoted as LWNet Stage II, the target Zernike coefficient set is denoted as Zerni ke Cofficients, and the target wave aberration is denoted as W2(x,y). The process of wave aberration reconstruction will be described as an example.

[0091] As Figure 6 shown, the process of wave aberration reconstruction includes the following steps:

[0092] Step 1: Take the actual point spread function image as the input, and select the first training network for supervised learning prediction to obtain the initial wave aberration W0(x,y).

[0093] Step 2: Project the initial wave aberration W0(x,y) into the Zernike coefficient space to obtain a set of Zernike coefficients

[0094] Step 3: Take Input the second training network, use the difference between the target point spread function and the actual point spread function as the loss function, and further optimize it through self-supervised learning to output the target Zernike coefficient set.

[0095] Step 4: Combine the target Zernike coefficient set into the target wave aberration W2(x, y).

[0096] Figure 7 It is a comparison schematic diagram of wave aberration reconstruction results obtained by different implementation schemes provided by an embodiment of the present invention. Among them, Figure 7 a in is a schematic diagram of the actual point spread function image and the corresponding actual wave aberration, b is a schematic diagram of the target point spread function image and the corresponding wave aberration reconstructed by the existing supervised method, and c is a schematic diagram of the target point spread function image and the corresponding wave aberration reconstructed by this scheme. As Figure 7 shown, this scheme has good effects in the wave aberration estimation of the actual imaging system, and the effects are always better than the reconstruction effects of the prior learning methods in the prior art.

[0097] In one embodiment, Figure 8 It is a schematic structural diagram of a wave aberration reconstruction device provided according to an embodiment of the present invention. This embodiment can execute the above-mentioned implementation manners. This embodiment is applicable to the situation of reconstructing wave aberration. The device can be implemented in a hardware / software manner and can be configured in an electronic device.

[0098] As Figure 8 shown, the wave aberration reconstruction device provided in this embodiment includes: an initial set acquisition module 401, a target set acquisition module 402, and a target wave aberration acquisition module 403, where:

[0099] The initial set acquisition module 401 is configured to obtain the corresponding initial Zernike coefficient set based on the pre-acquired actual point spread function image and the first training network;

[0100] The target set acquisition module 402 is configured to optimize the initial Zernike coefficient set by using the second training network based on the principle that the target point spread function image is the same as the actual point spread function image, and obtain the corresponding target Zernike coefficient set;

[0101] The target wave aberration acquisition module 403 is configured to obtain the corresponding target wave aberration based on the target Zernike coefficient set and the pre-configured optical system basis function.

[0102] Based on the principle that the target point spread function image is the same as the actual point spread function image, this embodiment measures the corresponding target point spread function image using the best imaging plane corresponding to the pre-configured optical system basis function, thus solving the problem of inconvenient operation caused by the need for multiple measurements and calibrations in wavefront aberration reconstruction, and simply and conveniently completing the process of wavefront aberration reconstruction, saving manpower and time, and improving the efficiency of wavefront aberration reconstruction; moreover, two training networks are used for wavefront aberration reconstruction, solving the problem of low accuracy in the prior art and improving the reconstruction accuracy.

[0103] Based on the above embodiment, the initial set acquisition module 401 includes:

[0104] An initial wavefront aberration acquisition unit, configured to input the actual point spread function image into the first training network to obtain the corresponding initial wavefront aberration.

[0105] An initial set acquisition unit, configured to project the initial wavefront aberration into the Zernike coefficient space to obtain the corresponding initial Zernike coefficient set.

[0106] Based on the above embodiment, the target set acquisition module 402 includes:

[0107] An intermediate coefficient acquisition unit, configured to adjust the initial weight coefficient of the second training network based on the initial Zernike coefficient set to obtain the corresponding intermediate weight coefficient.

[0108] An intermediate coefficient set acquisition unit, configured to obtain the corresponding intermediate Zernike coefficient set based on the second training network and the intermediate weight coefficient.

[0109] An image acquisition unit, configured to obtain the corresponding target point spread function image based on the intermediate Zernike coefficient set.

[0110] A target coefficient acquisition unit, configured to adjust the intermediate weight coefficient of the second training network based on the principle that the target point spread function image is the same as the actual point spread function image to obtain the corresponding target weight coefficient.

[0111] A target set acquisition unit, configured to obtain the corresponding target Zernike coefficient set based on the second training network and the target weight coefficient.

[0112] Based on the above embodiment, the image acquisition unit includes:

[0113] An intermediate wavefront aberration acquisition subunit, configured to perform a weighted linear combination based on the intermediate Zernike coefficient set and the pre-configured optical system basis function to obtain the corresponding intermediate wavefront aberration.

[0114] A function determination subunit, configured to determine the corresponding generalized pupil function based on the intermediate wavefront aberration, the pre-configured aperture function, and the wavelength.

[0115] A function image determination subunit, configured to use the square of the magnitude of the Fourier transform of the generalized pupil function as the corresponding target point spread function image.

[0116] Based on the above embodiments, the principle that the target point spread function image is the same as the actual point spread function image is specifically used for:

[0117] Determine the actual loss value between the target point spread function image and the actual point spread function image;

[0118] When the actual loss value reaches the minimum, determine that the target point spread function image is the same as the actual point spread function image.

[0119] Based on the above embodiments, the target wave aberration acquisition module 403 is specifically configured to perform a weighted linear combination of the target Zernike coefficient set and the pre-configured optical system basis function to obtain the corresponding target wave aberration.

[0120] Based on the above embodiments, the first training network is a supervised learning network; the second training network is a self-supervised learning network.

[0121] The wave aberration reconstruction device provided by the embodiments of the present invention can execute any wave aberration reconstruction method provided by the embodiments of the present invention, and has the corresponding functional modules and beneficial effects for executing the method. The content not described in detail in this embodiment can be referred to the description in any method embodiment of the present invention.

[0122] In one embodiment, Figure 9 It is a schematic structural diagram of an electronic device for implementing the wave aberration reconstruction method of the embodiments of the present invention. The electronic device 50 that can be used to implement the embodiments of the present invention is intended to represent various forms of digital computers, such as, laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as, personal digital processors, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are only examples and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0123] Such as Figure 9As shown, the electronic device 50 includes at least one processor 51 and a memory communicatively connected to the at least one processor 51, such as a read-only memory (ROM) 52, a random access memory (RAM) 53, etc. The memory stores a computer program executable by the at least one processor. The processor 51 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 52 or the computer program loaded from the storage unit 58 into the random access memory (RAM) 53. In the RAM 53, various programs and data required for the operation of the electronic device 50 can also be stored. The processor 51, the RAM 52, and the RAM 53 are connected to each other via a bus 54. An input / output (I / O) interface 55 is also connected to the bus 54.

[0124] Multiple components in the electronic device 50 are connected to the I / O interface 55, including: an input unit 56, such as a keyboard, a mouse, etc.; an output unit 57, such as various types of displays, speakers, etc.; a storage unit 58, such as a magnetic disk, an optical disc, etc.; and a communication unit 59, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 59 allows the electronic device 50 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0125] The processor 51 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the processor 51 include but are not limited to a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 51 executes the various methods and processes described above, such as the wavefront aberration reconstruction method.

[0126] In some embodiments, the wavefront aberration reconstruction method can be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as the storage unit 58. In some embodiments, part or all of the computer program can be loaded and / or installed onto the electronic device 50 via the ROM 52 and / or the communication unit 59. When the computer program is loaded into the RAM 53 and executed by the processor 51, one or more steps of the wavefront aberration reconstruction method described above can be executed. Alternatively, in other embodiments, the processor 51 can be configured to execute the wavefront aberration reconstruction method by any other appropriate means (e.g., by means of firmware).

[0127] The various embodiments of the systems and techniques described above in this specification can be implemented in digital electronic circuitry, integrated circuit systems, field programmable gate arrays (FPGA), application specific integrated circuits (ASIC), application specific standard products (ASSP), systems on a chip (SOC), complex programmable logic devices (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include: being implemented in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which may be a special-purpose or general-purpose programmable processor that receives data and instructions from, and transmits data and instructions to, a storage system, at least one input device, and at least one output device.

[0128] The computer program for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the computer programs, when executed by the processor, cause the functions / operations specified in the flowchart and / or block diagram to be implemented. The computer program may be executed entirely on the machine, partly on the machine, as a stand-alone software package partly on the machine and partly on a remote machine or entirely on the remote machine or server.

[0129] In the context of the present invention, a computer-readable storage medium may be a tangible medium that can contain, or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. The computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, the computer-readable storage medium may be a machine-readable signal medium. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0130] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or a trackball) through which the user can provide input to the electronic device. Other kinds of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0131] The systems and techniques described herein can be implemented in a computing system including backend components (e.g., as a data server), or a computing system including middleware components (e.g., an application server), or a computing system including frontend components (e.g., a user computer having a graphical user interface or a web browser through which the user can interact with an implementation of the systems and techniques described herein), or a computing system including any combination of such backend components, middleware components, or frontend components. The components of the system can be interconnected to each other by digital data communication in any form or medium (e.g., a communication network). Examples of communication networks include: local area network (LAN), wide area network (WAN), blockchain network, and the Internet.

[0132] The computing system can include a client and a server. The client and the server are generally remote from each other and typically interact through a communication network. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or a cloud host, which is a host product in the cloud computing service system and solves the defects of difficult management and weak business scalability existing in traditional physical hosts and VPS services.

[0133] It should be understood that various forms of the processes shown above can be used, with steps reordered, added, or deleted. For example, the steps recited in the present invention can be executed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution of the present invention can be achieved, and no limitation is imposed herein.

[0134] The above specific embodiments do not constitute a limitation on the protection scope of the present invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A wave aberration reconstruction method, characterized in that: include: Based on the actual point spread function image acquired in advance and the first training network, a corresponding initial Zernike coefficient set is obtained; Based on the principle that the target point spread function image is the same as the actual point spread function image, the initial Zernike coefficient set is optimized using a second training network to obtain a corresponding target Zernike coefficient set; Obtaining a corresponding target wave aberration based on the target Zernike coefficient set and a preconfigured optical system basis function; The method is based on the principle that the target point spread function image is the same as the actual point spread function image, and uses a second training network to optimize the initial Zernike coefficient set to obtain a corresponding target Zernike coefficient set, including: Adjusting the initial weight coefficients of the second training network based on the initial Zernike coefficient set to obtain corresponding intermediate weight coefficients; Obtaining a corresponding set of intermediate Zernike coefficients based on the second training network and the intermediate weight coefficients; Obtaining a corresponding target point spread function image based on the intermediate Zernike coefficient set; Based on the principle that the target point spread function image is the same as the actual point spread function image, adjusting the intermediate weight coefficient of the second training network to obtain a corresponding target weight coefficient; A corresponding target Zernike coefficient set is obtained based on the second training network and the target weight coefficient.

2. The method according to claim 1, characterized in that The method of obtaining a corresponding initial Zernike coefficient set based on the actual point spread function image acquired in advance and the first training network includes: Inputting the actual point spread function image into the first training network to obtain the corresponding initial wave aberration; The initial wave aberration is projected into the Zernike coefficient space to obtain a corresponding initial Zernike coefficient set.

3. The method according to claim 1, characterized in that The step of obtaining a corresponding target point spread function image based on the intermediate Zernike coefficient set includes: Performing a weighted linear combination based on the intermediate Zernike coefficient set and a pre-configured optical system basis function to obtain a corresponding intermediate wave aberration; Determining a corresponding generalized pupil function based on the intermediate wave aberration and a preconfigured aperture function and wavelength; The square of the magnitude of the Fourier transform of the generalized pupil function is used as the corresponding target point spread function image.

4. The method according to claim 1, characterized in that: The target point spread function image and the actual point spread function image have the same principles, including: Determining an actual loss value between the target point spread function image and the actual point spread function image; When the actual loss value reaches a minimum value, it is determined that the target point spread function image is the same as the actual point spread function image.

5. The method according to claim 1, characterized in that The obtaining of the corresponding target wave aberration based on the target Zernike coefficient set and the pre-configured optical system basis function comprises: A weighted linear combination is performed on the target Zernike coefficient set and a pre-configured optical system basis function to obtain a corresponding target wavefront aberration.

6. The method according to any one of claims 1 to 5, characterized in that: The first training network is a supervised learning network; the second training network is a self-supervised learning network.

7. A wave aberration reconstruction device, characterized in that: The device comprises: An initial set acquisition module, used to obtain a corresponding initial Zernike coefficient set based on a pre-acquired actual point spread function image and a first training network; A target set acquisition module, configured to optimize the initial Zernike coefficient set based on the principle that the target point spread function image is the same as the actual point spread function image and to obtain a corresponding target Zernike coefficient set by using a second training network; A target wave aberration acquisition module, used for obtaining a corresponding target wave aberration based on the target Zernike coefficient set and a pre-configured optical system basis function; The target set acquisition module includes: An intermediate coefficient acquisition unit, used to adjust the initial weight coefficient of the second training network based on the initial Zernike coefficient set to obtain a corresponding intermediate weight coefficient; An intermediate coefficient set acquisition unit, used to obtain a corresponding intermediate Zernike coefficient set based on the second training network and the intermediate weight coefficient; An image acquisition unit, used for obtaining a corresponding target point spread function image based on the intermediate Zernike coefficient set; A target coefficient acquisition unit, used to adjust the intermediate weight coefficient of the second training network based on the principle that the target point spread function image is the same as the actual point spread function image, so as to obtain the corresponding target weight coefficient; The target set acquisition unit is used to obtain a corresponding target Zernike coefficient set based on the second training network and the target weight coefficient.

8. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform the wave aberration reconstruction method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the wave aberration reconstruction method according to any one of claims 1 to 6 when executed.

Images