Multi-surface phase shift interference measurement method and system based on LSTM (Long Short Term Memory)

Abstract

The invention belongs to the field of optical precision measurement, and provides a multi-surface phase shift interference measurement method and system based on LSTM, and the method comprises the steps: placing a test mirror at a position alpha away from a reference mirror, carrying out the sampling of an interferometer light path according to a preset sampling time sequence during the wavelength tuning, and obtaining a sampling time sequence; calculating corresponding phases of the first interference cavity, the second interference cavity and the third interference cavity based on the central wavelength, and calculating light intensity of the receiving end of the interferometer based on the phases to obtain a light intensity matrix; inputting the light intensity matrix as a training data set and the phase as a label into an LSTM-based neural network for training, and obtaining a resolving network after the training is completed; and placing a to-be-measured mirror at a position alpha away from the reference mirror, collecting an interference pattern photo and converting the interference pattern photo into a gray matrix, inputting the gray matrix into the resolving network to obtain a phase value, and unwrapping the phase value to obtain a surface shape of the to-be-measured mirror. Compared with the prior art, the method for accurately reconstructing the surface shape under the non-ideal phase shift condition is provided.

Description

Technical Field

[0001] The present invention relates to the field of optical precision measurement, and in particular to a multi-surface phase shift interferometry measurement method and system based on LSTM. Background Art

[0002] Traditional phase-shift interferometry algorithms were designed in the early days to solve the surface shape solution of a single surface, and they often achieve phase shift by moving the reference plane. For surface shape detection of optical devices containing multiple nearly parallel surfaces, it is usually necessary to apply substances such as vaseline to non-measurement surfaces to scatter the reflected light from the non-measurement surfaces to avoid affecting the main interference signal. In view of this inconvenience, another phase-shift interferometry measurement method for multiple surfaces was proposed. In order to solve the problems of signal separation and accurate solution in multi-surface phase-shift interferometry, researchers have proposed a variety of improvement methods. For example, in the field of frequency domain analysis, the use of Fourier transform combined with window function can effectively separate interference signals in the frequency domain, thereby simplifying the complexity of equation solution, while enhancing harmonic suppression performance to reduce the impact of phase shift errors.

[0003] At present, in actual measurement, in order to facilitate the surface shape solution, it is usually required to achieve equal-step phase shift as much as possible, that is, to ensure that the laser wavelength modulation rate is stable, the image sampling is accurate, and the uniformity of the phase shift interval is met as much as possible to reduce the impact of the phase shift error. However, in actual operation, undesirable phase shift conditions are often encountered. For example, due to the nonlinear response of the equipment or environmental interference, the phase shift amount of each step may deviate from the expectation, showing a constant offset or random fluctuation. This deviation will cause the phase information in the interference pattern to be distorted, thereby affecting the final surface shape solution accuracy. Traditional phase shift interferometry calculation methods usually assume that the phase shift conditions are ideal, and lack effective error compensation capabilities for such deviations. Therefore, it is impossible to accurately reconstruct the surface shape under undesirable phase shift conditions. Summary of the invention

[0004] In order to overcome the defect of the above-mentioned prior art that the surface shape cannot be accurately reconstructed under non-ideal phase shift conditions, the present invention provides a multi-surface phase shift interferometry method and system based on LSTM.

[0005] In order to achieve the above technical effects, the technical solution of the present invention is as follows:

[0006] A multi-surface phase-shift interferometry method based on LSTM, comprising the following steps:

[0007] The test mirror is placed at a distance α from the reference mirror, and the interference light path passes through the reference mirror, the first plane of the test mirror and the second plane of the test mirror in sequence; a first interference cavity is formed between the reference mirror and the first plane of the test mirror; a second interference cavity is formed between the reference mirror and the second plane of the test mirror; and the first plane and the second plane of the test mirror form a third interference cavity;

[0008] When performing wavelength tuning, the interferometer optical path is sampled at a preset sampling timing, the phases corresponding to the first interferometer cavity, the second interferometer cavity and the third interferometer cavity are calculated based on the central wavelength, and the light intensity at the receiving end of the interferometer is calculated based on the phase to obtain a light intensity matrix;

[0009] The light intensity matrix is ​​used as a training data set, and the phase of the first interferometer cavity corresponding to the light intensity matrix is ​​used as a label to be input into a LSTM-based neural network for training, and a solution network is obtained after the training is completed;

[0010] The mirror to be tested is placed at a distance α from the reference mirror, an interference pattern photo is collected and converted into a grayscale matrix, the grayscale matrix is ​​input into the solution network to obtain a phase value, the phase value is unwrapped to obtain a surface diagram of the mirror to be tested.

[0011] The present invention also proposes a multi-surface phase shift interferometry system based on LSTM, the system comprising:

[0012] Training data acquisition module: samples the interferometer optical path at a preset sampling timing, and calculates the phase corresponding to the interferometer cavity and the light intensity on the receiving optical path;

[0013] Model training module: It is equipped with a LSTM-based neural network. The data collected in the training data collection module is input into the LSTM-based neural network for training. After the training is completed, the solution network is obtained;

[0014] Measuring module: collects interference pattern photos of the mirror to be measured and converts them into grayscale matrix, inputs the grayscale matrix into the solving network to obtain phase value, unpacks the phase value to obtain the surface figure of the mirror to be measured.

[0015] The present invention also proposes 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 LSTM-based multi-surface phase shift interferometry method as described in the present invention is implemented.

[0016] The present invention also proposes a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the multi-surface phase-shift interferometry method based on LSTM as described in the present invention is implemented.

[0017] Compared with the prior art, the present invention has the following beneficial effects:

[0018] The present invention learns the phase information of the interference pattern from the input data through the LSTM neural network. Since the random phase shift factor usually exists in different ways in the training samples, the trained model can generalize this change pattern and automatically compensate for these errors in the inference stage. Secondly, the LSTM neural network can capture the long-term trend from the grayscale change sequence, while the random phase shift factor usually exists in the form of noise, which is manifested as a short-term offset or disturbance. The LSTM neural network can ignore these short-term random fluctuations by capturing the main trend characteristics, so that the final phase solution is insensitive to the random phase shift factor. In addition, during the LSTM neural network training process, it directly learns the relationship between the phase and grayscale changes from the grayscale sequence without relying on the assumptions of the traditional method on the phase shift step or error. Therefore, when the network processes the interference pattern under the random phase shift condition, it can be directly mapped to the phase value through the learned weights without relying on the precise step or phase shift conditions as in the traditional method. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a flow chart of a multi-surface phase-shift interferometry method based on LSTM.

[0020] Figure 2 Diagram of the placement of reference mirrors and test mirrors.

[0021] Figure 3 is the wavelength response curve of the laser and the fitted response curve.

[0022] Figure 4 This is the architecture diagram of the LSTM-based neural network.

[0023] Figure 5 The following is an architecture diagram of a multi-surface phase-shift interferometry system based on LSTM.

[0024] Figure 6 This is the optical path of the Fizeau interferometer constructed in Example 6.

[0025] Figure 7 This is the surface measurement result of the multi-surface phase-shifting interferometry method based on LSTM.

[0026] Figure 8 This is the surface measurement result of the 45-step algorithm.

[0027] Fig. 9 This is the surface diagram obtained by measuring the present application and the 45-step algorithm under random phase shift conditions. DETAILED DESCRIPTION

[0028] The accompanying drawings are only used for illustrative purposes and are not to be construed as limiting the present invention;

[0029] It is understandable to those skilled in the art that some well-known descriptions may be omitted in the drawings.

[0030] The technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0031] Example 1

[0032] This embodiment proposes a multi-surface phase shift interferometry method based on LSTM, such as Figure 1 As shown, it is a flow chart of the multi-surface phase-shifting interferometry measurement method based on LSTM of this embodiment.

[0033] The multi-surface phase shift interferometry method based on LSTM proposed in this embodiment includes the following steps:

[0034] The test mirror is placed at a distance α from the reference mirror, and the interference light path passes through the reference mirror, the first plane of the test mirror and the second plane of the test mirror in sequence; wherein a first interference cavity is formed between the reference mirror and the first plane of the test mirror; a second interference cavity is formed between the reference mirror and the second plane of the test mirror; and a third interference cavity is formed between the first plane and the second plane of the test mirror;

[0035] When performing wavelength tuning, the interferometer optical path is sampled at a preset sampling timing, the phases corresponding to the first interferometer cavity, the second interferometer cavity and the third interferometer cavity are calculated based on the central wavelength, and the light intensity at the receiving end of the interferometer is calculated based on the phase to obtain a light intensity matrix;

[0036] The light intensity matrix is ​​used as a training data set, and the phase of the first interferometer cavity corresponding to the light intensity matrix is ​​used as a label to be input into a LSTM-based neural network for training, and a solution network is obtained after the training is completed;

[0037] The mirror to be tested is placed at a distance α from the reference mirror, an interference pattern photo is collected and converted into a grayscale matrix, the grayscale matrix is ​​input into the solution network to obtain a phase value, the phase value is unwrapped to obtain a surface diagram of the mirror to be tested.

[0038] In this embodiment, the phase information of the interference pattern is extracted from the input data using the LSTM neural network. Since the random phase shift factor exists in a variety of ways in the training sample, the trained model can effectively learn and generalize these change patterns, thereby automatically compensating for the resulting errors in the reasoning stage. In addition, the LSTM neural network has the ability to capture long-term trends in grayscale change sequences, while the random phase shift factor usually manifests itself as a short-term disturbance or offset. By focusing on the main trend features, the LSTM neural network can effectively ignore these short-term random fluctuations, thereby improving the robustness of the phase solution to the random phase shift factor. Further, during the training process of the LSTM neural network, the network directly learns the intrinsic relationship between phase and grayscale changes from the grayscale sequence without relying on the assumptions of the traditional method on the phase shift step or error. Therefore, when the model processes the interference pattern with random phase shift conditions, the input data can be directly mapped to the phase value through the learned weights, without relying on the precise step or specific phase shift conditions as in the traditional method.

[0039] As an exemplary illustration, the test mirror is a transparent parallel plate.

[0040] like Figure 2 As shown in FIG. 1 , it is a diagram of the placement of the reference mirror and the test mirror, wherein r is the reference mirror, u is the first plane of the test mirror, and d is the second plane of the test mirror.

[0041] Specifically, three interference cavities are generated by transparent parallel plates, and the second and third interference cavities can be regarded as interference signals. By using the phase information of the first interference cavity as a label, the LSTM neural network can directly learn the nonlinear mapping relationship between the input grayscale sequence and the target phase. Although the interference signals generated by the second and third interference cavities will affect the overall interference pattern, through the end-to-end learning capability of the LSTM neural network, the neural network can identify and ignore these minor interference signals and focus on the phase characteristics of the first interference cavity to ensure the accuracy and reliability of the measurement results. Through this design, the system can adapt to a variety of configurations of different test mirrors such as transparent parallel plates, and is suitable for complex optical measurement scenarios.

[0042] In an optional embodiment, the step of sampling the interferometer optical path at a preset sampling timing includes: performing fourth-order polynomial fitting on the wavelength response curve of the laser, and sampling the fitted curve at a preset sampling timing.

[0043] like Figure 3 Shown are the wavelength response curve of the laser and the fitted response curve.

[0044] In this embodiment, the fitting operation is equivalent to converting the discrete signal into a continuous signal approximately. The high-order phase shift error comes from the nonlinearity of laser wavelength tuning. Obtaining sampling points on the fitted response curve is equivalent to taking the nonlinear factor into account in the training set, eliminating the influence of the high-order phase shift error from the root.

[0045] In an optional embodiment, the method further includes: changing the cavity length of the interference cavity, and calculating the corresponding phase and the light intensity on the test mirror under different cavity lengths to obtain a number of light intensity matrices, using the several light intensity matrices as training data, and the first interference cavity length corresponding to any matrix as a label input into the LSTM-based neural network for training.

[0046] Specifically, the interference signal intensity I(x,y) received at a certain pixel point in the imaging system of the interferometer can be expressed as:

[0047]

[0048] Wherein, I0(x, y) is the background light intensity, I1, I2 and I3 are the modulation amplitudes of the interference signals generated by the first interferometer cavity, the second interferometer cavity and the third interferometer cavity respectively; φ1(x, y), φ2(x, y) and φ3(x, y) are the initial phases of the three groups of interference signals respectively.

[0049] φ1(x,y)=2kL ru (x,y)

[0050] φ2(x,y)=2k[L ru (x,y)+nL rd (x,y)]

[0051] φ3(x,y)=2knL ud (x,y)

[0052] Among them, L ru is the first interference cavity, L rd is the second interference cavity, L ud is the third interference cavity; k=2π / λ, n is the refractive index of the material to be measured.

[0053] In this embodiment, by changing the cavity length of the interferometer cavity and calculating the corresponding phase under different cavity lengths and the light intensity on the test mirror, the diversity of the training data can be significantly increased, so that the training set data is greatly increased, covering a wider range of phase distribution characteristics, thereby improving the generalization ability and robustness of the neural network model. This method not only reduces the dependence on real data collection, effectively reduces experimental costs and time expenses, but also enhances the model's adaptability to complex environments and random noise.

[0054] Further optionally, the step of changing the length of the interference cavity includes: changing the undulation of the test mirror surface; the calculation formula of the cavity length of the interference cavity is as follows:

[0055] L ru (x,y)=L1(x,y)+Z u0 (x,y)

[0056] L rd (x,y)=L2(x,y)+Z d0 (x,y)

[0057] L ud (x,y)=L2(x,y)-L1(x,y)

[0058] Where (x, y) represents any point on the test mirror, L1 is the distance between the first plane of the test mirror and the reference mirror, L2 is the distance between the second plane of the test mirror and the reference mirror, and Z u0 (x, y) is the undulation of the first plane of the test mirror, Z d0 (x, y) is the fluctuation of the second plane of the test mirror, L ru , L rd and L ud They are respectively the first interference cavity, the second interference cavity and the third interference cavity.

[0059] As an example, Z u0 (x,y) and Z d0 The magnitude of (x,y) is from nanometer to micrometer.

[0060] In an optional embodiment, the changing of the undulation of the test mirror surface includes: adjusting the undulation of the first plane of the test mirror and the undulation of the second plane of the test mirror within a preset range and arbitrarily arranging and combining them.

[0061] Specifically, Z u0 (x,y) and Z d0 (x, y) takes values ​​within a preset range and arranges and combines them arbitrarily to obtain several combinations of the three initial phase values ​​φ1(x, y), φ2(x, y), and φ3(x, y), and calculate the corresponding light intensity. The sequence of phase φ1(x, y) is taken as the remainder of 2π, and the remainders are numbered in order of size. The numbers are input into the neural network as simplified label values.

[0062] In this embodiment, by changing the fluctuation of the test mirror surface, the cavity length of the interferometer cavity can be flexibly adjusted on a microscopic scale, thereby generating diversified phase data under different cavity length conditions, significantly increasing the richness of the training data and covering a wider phase distribution range. In addition, the generated microscopic fluctuation changes can also simulate the characteristics of the interferometer cavity under a variety of actual measurement scenarios, making the neural network more adaptable, and still maintaining high-precision phase solution performance when processing interference patterns in complex or random environments, providing a guarantee for the accuracy and stability of optical measurements.

[0063] Further optionally, the LSTM-based neural network contains several LSTM layers, each LSTM layer contains multiple neurons, and an activation layer is placed after each neuron.

[0064] like Figure 4 As shown, this is the architecture diagram of the LSTM-based neural network.

[0065] In this embodiment, deep features are extracted layer by layer through a multi-layer LSTM structure, and the temporal correlation and nonlinear relationship in the interference pattern data are modeled more accurately. At the same time, the introduction of the activation layer alleviates the gradient vanishing problem, optimizes the gradient flow and training stability of the network, thereby accelerating convergence and improving training efficiency.

[0066] Example 2

[0067] This embodiment makes improvements on the LSTM-based multi-surface phase-shifting interferometry measurement method proposed in Embodiment 1.

[0068] In an optional embodiment, the method further includes: when using laser wavelength tuning for phase shifting, introducing a number of constant phase shift offsets, and inputting the phase and light intensity calculated based on each constant phase shift offset into a LSTM-based neural network for training, and obtaining a number of solution networks after the training is completed; inputting the serial numbers of the several solution networks as labels into a LSTM-based neural network for training to obtain a judgment network; the judgment network measures the phase shift offset based on the grayscale change matrix of the mirror to be measured, and selects a suitable solution network according to the size of the phase shift offset.

[0069] In this embodiment, by introducing several constant phase shift offsets, the coverage of the training data can be effectively expanded, and diversified data sets under different offset conditions can be generated, thereby improving the adaptability and robustness of the neural network to the constant offset. Furthermore, the several solution networks obtained through training have more targeted solution performance than a single network model, and can significantly improve the accuracy and efficiency of phase solution. In addition, by judging the network, it can select the solution network that best suits the current environment and specific conditions according to the gray matrix characteristics of the mirror to be measured, thereby improving the flexibility of the system in complex measurement scenarios and optimizing the stability and accuracy of the solution process.

[0070] In an optional embodiment, the step of introducing a plurality of constant phase shift offsets includes: presetting a plurality of constant phase shift offsets, changing the sampling interval of the sampling sequence according to the magnitude of the phase shift offsets, and obtaining the light intensity matrix and phase corresponding to different phase shift offsets.

[0071] As an example, set the time interval to The interference light path was sampled 50 times for 1 second to obtain the initial light intensity matrix. The sampling time interval was then increased or decreased with a step size of 1% and an upper and lower limit of ±25% to obtain another 50 different time intervals, and 50 samples were taken for each of them. A total of 51 light intensity matrices with different constant phase shift offsets were obtained.

[0072] In this embodiment, through time interval sampling, multiple constant offset phase shift conditions can be dynamically generated on the time axis to enrich the diversity of the training data set, thereby covering more diverse phase and light intensity change characteristics. Furthermore, by combining the multi-solution network and the judgment network, strong support is provided for the flexibility and reliability of this method in complex optical measurement tasks.

[0073] Example 3

[0074] This embodiment proposes a multi-surface phase shift interferometry system based on LSTM, and applies the multi-surface phase shift interferometry method based on LSTM proposed in Embodiments 1 and 2. Figure 5 , which is an architecture diagram of the LSTM-based multi-surface phase-shift interferometry measurement system of this embodiment.

[0075] This embodiment proposes a multi-surface phase shift interferometry measurement system based on LSTM, including:

[0076] Training data acquisition module: samples the laser light path in time and calculates the phase corresponding to the interferometer cavity and the light intensity on the test mirror;

[0077] Model training module: It is equipped with a LSTM-based neural network. The data collected in the training data collection module is input into the LSTM-based neural network for training. After the training is completed, the solution network is obtained;

[0078] Measuring module: collects interference pattern photos of the mirror to be measured and converts them into grayscale matrix, inputs the grayscale matrix into the solving network to obtain phase value, unpacks the phase value to obtain the surface figure of the mirror to be measured.

[0079] It can be understood that the system of this embodiment corresponds to the method of the above-mentioned embodiment 1 and embodiment 2, and the options in the above-mentioned embodiment 1 and embodiment 2 are also applicable to this embodiment, so they are not described repeatedly here.

[0080] Example 4

[0081] This embodiment proposes a computer device, including a memory and a processor, wherein the memory stores computer-readable instructions, wherein when the computer-readable instructions are executed by the processor, the processor executes the steps of the LSTM-based multi-surface phase-shift interferometry measurement method proposed in Embodiments 1 and 2.

[0082] Example 5

[0083] This embodiment proposes a storage medium on which computer-readable instructions are stored, wherein the computer-readable instructions, when executed by a processor, implement the steps of the LSTM-based multi-surface phase-shift interferometry measurement method proposed in Embodiments 1 and 2.

[0084] Exemplarily, the storage medium includes, but is not limited to, a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0085] Exemplarily, the instructions, programs, code sets or instruction sets may be implemented using conventional programming languages.

[0086] Exemplarily, the processor includes but is not limited to a smart phone, a personal computer, a server, a network device, etc., and is used to execute all or part of the steps of the LSTM-based multi-surface phase-shift interferometry measurement method based on staged learning described in Examples 1 and 2.

[0087] Example 6

[0088] This embodiment applies a multi-surface phase-shift interferometry measurement method based on LSTM proposed in Embodiments 1 to 5 to measure the mirror to be measured.

[0089] like Figure 6 As shown, the optical path of the Fizeau interferometer constructed in this embodiment.

[0090] like Figure 7 As shown, the surface measurement results of the multi-surface phase-shifting interferometry method based on LSTM.

[0091] like Figure 8 As shown, it is the surface measurement result of the 45-step algorithm.

[0092] The 45-step algorithm is the Characteristic Polynomial of The 45-sample Phase-shifting Algorithm, which is a phase extraction technique used in interferometry. The algorithm combines Fourier transform and phase shift techniques to accurately calculate phase information by collecting interference patterns in multiple step phases. Figure 7 and Figure 8 It can be seen that the measurement results of this method are close to those of the 45-step algorithm and can capture the main three-dimensional morphological features of the object surface.

[0093] Then, the linear phase shift factor ε0 and the random phase shift factor M are introduced respectively. As shown in Table 1 below, the surface PV value and RMS value of the present application and the 45-step solution results under different linear phase shifts ε0 are shown; a smaller PV value indicates a smaller surface morphology or wavefront error, and a smaller RMS value indicates a lower overall level of surface or wavefront error, which is smoother or closer to the ideal state as a whole. As can be seen from Table 1, the present application has excellent insensitivity in the range of ε0 = 0% to -25%. As ε0 increases, the PV value and RMS value of the 45-step algorithm increase, while the solution of this paper remains relatively stable.

[0094] Table 1 PV and RMS values ​​under linear phase shift conditions

[0095]

[0096] As shown in Table 2 below, the PV value and RMS value of the surface shape of this application and the 45-step solution results under different random phase shift factors M. As can be seen from Table 1, as M increases, the PV value of the 45-step algorithm gradually increases, while the calculated value of this solution remains relatively stable. In addition, this application can also maintain stability in RMS value.

[0097] Table 2 PV and RMS values ​​under random phase shift conditions

[0098]

[0099] like Fig. 9 As shown, the surface images measured by the present application and the 45-step algorithm under the condition of random phase shift, among which, ad is the surface image measured by the present application method, and eh is the surface image measured by the 45-step algorithm. It can be seen that as the amplitude of the random phase shift increases, the surface image obtained by the 45-step algorithm gradually becomes distorted, while the surface image measured by the present application still has the main three-dimensional morphological features.

[0100] The terms in the drawings are for illustrative purposes only and should not be construed as limiting this patent;

[0101] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the embodiments here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. A multi-surface phase shift interferometry method based on LSTM, characterized in that: The following steps are involved: The test mirror is placed at a distance α from the reference mirror, and the interference light path passes through the reference mirror, the first plane of the test mirror and the second plane of the test mirror in sequence; wherein a first interference cavity is formed between the reference mirror and the first plane of the test mirror; a second interference cavity is formed between the reference mirror and the second plane of the test mirror; and a third interference cavity is formed between the first plane and the second plane of the test mirror; When performing wavelength tuning, the interferometer optical path is sampled at a preset sampling timing, the phases corresponding to the first interferometer cavity, the second interferometer cavity and the third interferometer cavity are calculated based on the central wavelength, and the light intensity at the receiving end of the interferometer is calculated based on the phase to obtain a light intensity matrix; The light intensity matrix is ​​used as a training data set, and the phase of the first interferometer cavity corresponding to the light intensity matrix is ​​used as a label to be input into a LSTM-based neural network for training, and a solution network is obtained after the training is completed; The mirror to be tested is placed at a distance α from the reference mirror, an interference pattern photo is collected and converted into a grayscale matrix, the grayscale matrix is ​​input into the solution network to obtain a phase value, the phase value is unwrapped to obtain a surface diagram of the mirror to be tested.

2. The multi-surface phase shift interferometry measurement method based on LSTM according to claim 1, characterized in that: The step of sampling the interferometer optical path at a preset sampling sequence includes: performing fourth-order polynomial fitting on the wavelength response curve of the laser, and sampling the fitted curve at a preset sampling sequence.

3. The multi-surface phase shift interferometry method based on LSTM according to claim 1, characterized in that: The method also includes: changing the cavity length of the interference cavity, and calculating the corresponding phase under different cavity lengths and the light intensity on the test mirror to obtain a plurality of light intensity matrices, using the plurality of light intensity matrices as training data, and inputting the first interference cavity cavity length corresponding to any matrix as a label into the LSTM-based neural network for training.

4. The multi-surface phase shift interferometry method based on LSTM according to claim 3, characterized in that: The step of changing the length of the interference cavity includes: changing the fluctuation of the test mirror surface; the calculation formula of the cavity length of the interference cavity is as follows: L ru (x,y)=L1(x,y)+Z u0 (x,y) L rd (x,y)=L2(x,y)+Z d0 (x,y) L ud (x,y)=L2(x,y)-L1(x,y) Where (x, y) represents any point on the test mirror, L1 is the distance between the first plane of the test mirror and the reference mirror, L2 is the distance between the second plane of the test mirror and the reference mirror, and Z u0 (x, y) is the undulation of the first plane of the test mirror, Z d0 (x, y) is the fluctuation of the second plane of the test mirror, L ru , L rd and L ud They are respectively the first interference cavity, the second interference cavity and the third interference cavity.

5. The multi-surface phase shift interferometry method based on LSTM according to claim 4, characterized in that: The changing of the undulation of the test mirror surface includes: adjusting the undulation of the first plane of the test mirror and the undulation of the second plane of the test mirror within a preset range and arbitrarily arranging and combining them.

6. The multi-surface phase shift interferometry method based on LSTM according to any one of claims 1 to 5, characterized in that: The method further comprises: when performing phase shift using laser wavelength tuning, introducing a plurality of constant phase shift offsets, calculating phase and light intensity based on any constant phase shift offset, constructing a neural network based on LSTM for any constant phase shift offset, and inputting the phase and light intensity for training, and obtaining a plurality of solution networks after the training is completed; Constructing a judgment network based on LSTM, and inputting the serial numbers of several solution networks as labels into the judgment network for training; The judgment network measures the phase shift offset based on the grayscale matrix of the mirror to be measured, and selects a suitable solution network according to the size of the phase shift offset.

7. The multi-surface phase shift interferometry method based on LSTM according to claim 6, characterized in that: The step of introducing a plurality of constant phase shift offsets comprises: presetting a plurality of constant phase shift offsets, changing the sampling interval of the sampling sequence according to the magnitude of the phase shift offsets, and obtaining the light intensity matrix and phase corresponding to different phase shift offsets.

8. A multi-surface phase shift interferometry system based on LSTM, applied to the multi-surface phase shift interferometry method based on LSTM according to any one of claims 1 to 7, characterized in that: The system comprises: Training data acquisition module: When performing wavelength tuning, the interferometer optical path is sampled at a preset sampling timing, and the phase corresponding to the interferometer cavity and the light intensity on the receiving optical path are calculated; Model training module: It is equipped with a LSTM-based neural network. The data collected in the training data collection module is input into the LSTM-based neural network for training. After the training is completed, the solution network is obtained; Measuring module: collects interference pattern photos of the mirror to be measured and converts them into grayscale matrix, inputs the grayscale matrix into the solving network to obtain phase value, unpacks the phase value to obtain the surface figure of the mirror to be measured.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the LSTM-based multi-surface phase-shifting interferometry method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the LSTM-based multi-surface phase-shifting interferometry method according to any one of claims 1 to 7 is implemented.

Images