A computer-aided assembly method based on cross-screening theory

Abstract

The technical solution of the present invention is to provide a computer-aided alignment method based on cross-screening theory. This method addresses the issues of sensitivity matrix error positioning bias and absolute reliance on wavefront detection equipment that occur in traditional computer-aided alignment. This method, which requires no fitting and can utilize multiple assembly error information image carriers, seeks the initial error model that best matches the variation pattern of test results. This method is applicable to a variety of optical path configurations, including refractive, reflective, on-axis, and off-axis configurations, effectively reducing the difficulty of precise alignment of optical systems and improving their efficiency.

Description

Technical Field

[0001] The present invention relates to an optical system assembly and adjustment method, in particular to a computer-aided assembly and adjustment method based on cross-screening theory. Background Art

[0002] With the rapid development of optical technology, optical systems are becoming increasingly sophisticated, placing higher demands on assembly and adjustment accuracy. Traditional optical system assembly and adjustment relies primarily on manual experience, lacking a precise accuracy range and resulting in low assembly efficiency. This has led to the development of computer-aided assembly technology, which relies on numerical analysis of optical system imaging information to resolve assembly errors.

[0003] Most existing computer-aided alignment methods calculate assembly errors based on the sensitivity matrix method. This method approximates the fitted wavefront aberration Zernike coefficients as a linear combination of the misalignments of the individual optical components. The coefficient matrix of this linear equation is the sensitivity matrix. Due to the relative compensation between optical components, one set of fitted wavefront aberrations may correspond to multiple sets of coefficient matrix solutions. This makes the sensitivity matrix method susceptible to positioning errors in the error sources. Furthermore, this method is extremely dependent on wavefront detection information and the required test equipment, which greatly limits its scope of application. Summary of the Invention

[0004] The technical problem to be solved by the present invention is that the existing computer-aided alignment method uses a sensitivity matrix method, which results in positioning deviation of the error source and relies on wavefront detection equipment, which has a high threshold for use.

[0005] In order to solve the above technical problems, the technical solution of the present invention is to provide a computer-aided assembly method based on cross-screening theory, which is characterized by comprising the following steps:

[0006] Step 1: Determine the splitting method of the optical system based on the auxiliary assembly complexity attribute, and split the optical system into k groups of subsystems, where k is greater than or equal to 1 and less than n, where n is the total number of optical components;

[0007] For each optical subsystem obtained by splitting in step 1, the following steps are used to process it:

[0008] Step 2: Use optical design software to perform a large number of misaligned assembly simulations on the designed optical subsystem, and exhaustively obtain a large number of misaligned optical system wavefront image information F according to the preset accuracy range and step size. 1i Or misaligned optical system star point image information P 1i , wavefront image information F of each misaligned optical system 1i Or misaligned optical system star point image information P 1i Corresponding to an offset error model, all offset error models form the offset set M1, where F 1irepresents the wavefront image information of the i-th misaligned optical system, P 1i represents the star point image information of the i-th misaligned optical system, 1<i≤h, h is the exhaustive total number;

[0009] Step 3: Roughly assemble the optical subsystem and measure the wavefront image information F of the optical system t1 Or optical system star point image information P t1 ;

[0010] Step 4: The optical system wavefront image information F obtained in step 3 is t1 Or optical system star point image information P t1 The wavefront image information F of the misaligned optical system obtained in step 1 1i Or misaligned optical system star point image information P 1i Perform similarity matching and select some misalignment error models corresponding to the misaligned optical system wavefront image or misaligned optical system star point image with higher similarity from the misalignment set M1. These misalignment error models constitute the initial error model parent set M2, M2∈M1;

[0011] Step 5: Accurately adjust one or more error variables in the rough optical subsystem. The error variable adjustment value is ∆ xs , and measure again to obtain the changed optical system wavefront image information F t2 Or optical system star point image information P t2 ;

[0012] Step 6: Set the error variable adjustment value ∆ xs Feedback input is given to each misalignment error model in the initial error model parent set M2, and the misalignment assembly simulation is performed on the initial error model parent set M2 again to obtain the secondary simulation result of the initial error model parent set M2, that is, the wavefront image information F of the misaligned optical system of the secondary simulation is obtained. 2i Or misaligned optical system star point image information P 2i ;

[0013] Step 7: The optical system wavefront image information F obtained in step 5 is t2 Or optical system star point image information P t2 The wavefront image information F of the misaligned optical system obtained in step 6 2i Or misaligned optical system star point image information P 2i Perform similarity matching and select the wavefront image information F of the misaligned optical system with higher similarity from the initial error model parent set M2 2i Or misaligned optical system star point image information P 2i The corresponding partial imbalance error model, this part of the imbalance error model constitutes the subset M3 of the initial error model parent set, M3∈M2, completing a round of cross screening;

[0014] Step 8: Repeat steps 5, 6, and 7 to continuously cross-screen the subsets of the initial error model parent set until the final set of misaligned error models m is screened out. q , q is the number of cross-screening times; in each round of cross-screening, the error variables adjusted in the next round of cross-screening are different from those in the previous round of cross-screening, ensuring that each error variable is changed only once during the entire optical subsystem assembly and adjustment process;

[0015] Step 9: The final set of offset error models m is selected q The misalignment error model best conforms to the changing law of the wavefront or star point when the optical subsystem is adjusted. It corresponds to the initial misalignment of the optical system. The adjustment of the optical subsystem can be completed by adjusting according to this initial misalignment.

[0016] Preferably, in step 1, if the auxiliary assembly complexity attribute is less than the preset complexity threshold, the optical system is not split; if the auxiliary assembly complexity attribute is greater than the preset complexity threshold, the optical system is split to reduce the computing pressure, and the value of the number of split groups ensures that the assembly complexity attribute of each group of subsystems obtained by the split is less than the preset complexity threshold.

[0017] Preferably, in step 4 and step 7, parts with higher similarity are screened out by a preset screening ratio. Setting a higher screening ratio can ensure the accuracy of the desired result, while setting a lower screening ratio can accelerate the convergence speed.

[0018] Preferably, in step 9, according to the offset error model m q Interpolation calculations are performed to find a closer approximate solution to obtain a closer model of the offset error.

[0019] The present invention addresses the problems of sensitivity matrix error positioning deviation and absolute dependence on wavefront detection equipment in traditional computer-aided alignment, and provides a computer-aided alignment method that does not require fitting and can utilize a variety of assembly error information image carriers to find the initial error model that is closest to the variation law of the test results. The method is applicable to a variety of optical path structures such as refraction, reflection, coaxial, and off-axis, effectively reducing the difficulty of precise alignment of optical systems and improving the efficiency of optical system alignment. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 A flowchart of a computer-aided assembly and adjustment method based on cross-screening theory provided by the present invention;

[0021] Figure 2 Schematic diagram of the test wavefront image information F t1 ;

[0022] Figure 3Schematic diagram of the test wavefront image information F t2 ;

[0023] Figure 4 The error model type setting table is shown. DETAILED DESCRIPTION

[0024] The sizes and proportions shown in the drawings in this specification are for illustrative purposes only and are provided to facilitate the description. They are not intended to limit the conditions for implementation of the present invention and do not affect the efficacy of the present invention. The positional relationships of "upper," "lower," "inner," and "outer" in this specification are provided for convenience only and are not intended to limit the scope of implementation of the present invention. Any changes in these relative relationships without substantial changes in the technical content are considered to be within the scope of implementation of the present invention.

[0025] The present invention provides a computer-aided alignment method based on cross-screening theory. This method can utilize simulation results of various types of assembly error information, such as wavefront and star points, to assist in alignment. This method avoids the problem of positioning error sources caused by the sensitivity matrix method, while also avoiding dependence on wavefront detection equipment, thereby lowering the threshold for use. The method specifically includes the following steps:

[0026] Step 1: Determine the optical system splitting method based on the auxiliary assembly complexity attribute, where the auxiliary assembly complexity attribute includes the type and number of error variables of the optical components, the accuracy range and step size of the misaligned assembly simulation; the optical system splitting method is to split the optical system into k groups of subsystems, where k is greater than or equal to 1 and less than n, and n is the total number of optical components. When k=1, it means that the optical system will not be split. If the auxiliary assembly complexity attribute is less than the preset complexity threshold, the optical system will not be split. If the auxiliary assembly complexity attribute is greater than the preset complexity threshold, the optical system will be split to reduce computational pressure. The value of the number of split groups must ensure that the assembly complexity attribute of each subsystem obtained by splitting is less than the preset complexity threshold.

[0027] The optical system of this embodiment is designed as a lens group consisting of lens 1 and lens 2. It is assumed that this lens group has good coaxiality and only has an optical axis axial error. Among them, the error of lens 1 is -23μm and the error of lens 2 is +45μm. This lens group has few independent variables, so this embodiment does not split the optical system.

[0028] For each optical subsystem obtained by splitting in step 1, the following steps are used to process it.

[0029] Step 2: Use optical design software to perform a large number of misaligned assembly simulations on the designed optical subsystem, and exhaustively obtain a large number of misaligned optical system wavefront image information F according to the preset accuracy range and step size. 1i Or misaligned optical system star point image information P1i , wavefront image information F of each misaligned optical system 1i Or misaligned optical system star point image information P 1i Corresponding to an offset error model, all offset error models form the offset set M1, where F 1i represents the wavefront image information of the i-th misaligned optical system, P 1i Represents the star point image information of the i-th misaligned optical system, 1<i≤h, h is the exhaustive total number.

[0030] This embodiment uses Zemax software for simulation. The preset accuracy range of lens 1 and lens 2 is plus or minus 50 μm, and the step size is 25 μm, that is, the error model types of the lens group are 25, such as Figure 4 The error model type setting is shown. In this embodiment, the wavefront image information of the optical system is used as the error information carrier, so 25 types of wavefronts are obtained by exhaustive enumeration, which are defined as the misalignment set M1, which includes misalignment error models m1 to m2. 25 , corresponding to the wavefront image information F of the misaligned optical system 101 to F 125 .

[0031] Step 3: Roughly assemble the optical subsystem and measure the wavefront image information F of the optical system t1 Or optical system star point image information P t1 .

[0032] In this embodiment, after the optical system is roughly assembled, the wavefront image information F of the optical system is measured. t1 like Figure 2 shown.

[0033] Step 4: The optical system wavefront image information F obtained in step 3 is t1 Or optical system star point image information P t1 The wavefront image information F of the misaligned optical system obtained in step 1 1i Or misaligned optical system star point image information P 1i Perform similarity matching (using algorithms such as sum of squared errors, hashing, and SSIM) to select the partial misalignment error models corresponding to the misaligned optical system wavefront images or misaligned optical system star point images with high similarity from the misalignment set M1. These partial misalignment error models constitute the initial error model superset M2, where M2∈M1. A preset percentage can be used to select the portion with high similarity. A higher percentage ensures the accuracy of the desired result, while a lower percentage accelerates convergence.

[0034] In this embodiment, the optical system wavefront image information F t1 Wavefront image information F of misaligned optical system 101 to F125 Perform similarity matching, use the wavefront text to calculate the error variance value, and screen out some of the misalignment error models corresponding to the misaligned optical system wavefront image information with high similarity. These misalignment error models constitute the initial error model parent set M2. In this embodiment, the top 16% are screened out, and the screened misaligned optical system wavefront image information result is F 107 、F 114 、F 117 、F 124 , corresponding to the offset error models m7 and m 14 、m 17 、m 24 .

[0035] Step 5: Accurately adjust one or more error variables in the rough optical subsystem. The error variable adjustment value is ∆ xs , and measure again to obtain the changed optical system wavefront image information F t2 Or optical system star point image information P t2 The error variables include X-axis eccentricity, X-axis tilt, X-axis defocus, Y-axis eccentricity, Y-axis tilt, and Y-axis defocus. The adjustment value of the error variable is ∆ xs Should be larger than the preset step size.

[0036] In this embodiment, the axial position of the lens 2 in the rough optical system is adjusted by 50 μm in the negative direction, and the following is measured again: Figure 3 The optical system wavefront image information F shown t2 .

[0037] Step 6: Set the error variable adjustment value ∆ xs Feedback input is given to each misalignment error model in the initial error model parent set M2, and the misalignment assembly simulation is performed on the initial error model parent set M2 again to obtain the secondary simulation result of the initial error model parent set M2, that is, the wavefront image information F of the misaligned optical system of the secondary simulation is obtained. 2i Or misaligned optical system star point image information P 2i .

[0038] Step 7: The optical system wavefront image information F obtained in step 5 is t2 Or optical system star point image information P t2 The wavefront image information F of the misaligned optical system obtained in step 6 2i Or misaligned optical system star point image information P 2i Perform similarity matching and select the wavefront image information F of the misaligned optical system with higher similarity from the initial error model parent set M2 2i Or misaligned optical system star point image information P 2iThe corresponding partial imbalance error model constitutes the subset M3 of the initial error model parent set, M3∈M2, completing a round of cross screening.

[0039] Step 8: Repeat steps 5, 6, and 7 to continuously cross-screen the subsets of the initial error model parent set until the final set of misaligned error models m is screened out. q , q is the number of cross-screening rounds. During this process, the error variables of different rounds are independent of each other. The error variables adjusted in the next cross-screening round cannot be the same as those in the previous round, ensuring that each error variable changes only once during the entire optical subsystem assembly and adjustment process.

[0040] Step 9: The final set of offset error models m is selected q The misalignment error model best matches the wavefront variation law of the optical subsystem during adjustment. The misalignment error model corresponds to the initial misalignment of the optical system. Adjustment can be completed by adjusting according to this initial misalignment. q It is a discrete result screened from the offset set M1. In order to obtain a closer offset error model, the offset error model m q Interpolation calculations are performed to find a closer approximate solution.

[0041] Since the number of samples in this embodiment is small, the best error model has been obtained after one round of screening, so no loop is entered. The final set of offset error models m 14 The error parameters that best match the wavefront variation of this optical system during adjustment are (-25μm, 50μm). After error adjustment based on these error parameters, the error between Lens 1 and Lens 2 is corrected to (2μm, -5μm), completing computer-aided alignment. For higher-precision error correction, a smaller step size can be used for interpolation calculations, and a more accurate solution can be obtained through simulation and screening of the interpolation error model.

Claims

1. A computer-aided assembly method based on cross-screening theory, characterized in that: The following steps are involved: Step 1: Determine the splitting method of the optical system based on the auxiliary assembly complexity attribute, and split the optical system into k groups of subsystems, where k is greater than or equal to 1 and less than n, where n is the total number of optical components; For each optical subsystem obtained by splitting in step 1, the following steps are used to process it: Step 2: Use optical design software to perform a large number of misaligned assembly simulations on the designed optical subsystem, and exhaustively obtain a large number of misaligned optical system wavefront image information F according to the preset accuracy range and step size. 1i Or misaligned optical system star point image information P 1i , wavefront image information F of each misaligned optical system 1i Or misaligned optical system star point image information P 1i Corresponding to an offset error model, all offset error models form the offset set M1, where F 1i represents the wavefront image information of the i-th misaligned optical system, P 1i represents the star point image information of the i-th misaligned optical system, 1<i≤h, h is the exhaustive total number; Step 3: Roughly assemble the optical subsystem and measure the wavefront image information F of the optical system t1 Or optical system star point image information P t1 ; Step 4: The optical system wavefront image information F obtained in step 3 is t1 Or optical system star point image information P t1 The wavefront image information F of the misaligned optical system obtained in step 1 1i Or misaligned optical system star point image information P 1i Perform similarity matching and select some misalignment error models corresponding to the misaligned optical system wavefront image or misaligned optical system star point image with higher similarity from the misalignment set M1. These misalignment error models constitute the initial error model parent set M2, M2∈M1; Step 5: Accurately adjust one or more error variables in the rough optical subsystem. The error variable adjustment value is ∆ xs , and measure again to obtain the changed optical system wavefront image information F t2 Or optical system star point image information P t2 ; Step 6: Set the error variable adjustment value ∆ xs Feedback input is given to each misalignment error model in the initial error model parent set M2, and the misalignment assembly simulation is performed on the initial error model parent set M2 again to obtain the secondary simulation result of the initial error model parent set M2, that is, the wavefront image information F of the misaligned optical system of the secondary simulation is obtained. 2i Or misaligned optical system star point image information P 2i ; Step 7: The optical system wavefront image information F obtained in step 5 is t2 Or optical system star point image information P t2 The wavefront image information F of the misaligned optical system obtained in step 6 2i Or misaligned optical system star point image information P 2i Perform similarity matching and select the wavefront image information F of the misaligned optical system with higher similarity from the initial error model parent set M2 2i Or misaligned optical system star point image information P 2i The corresponding partial imbalance error model, this part of the imbalance error model constitutes the subset M3 of the initial error model parent set, M3∈M2, completing a round of cross screening; Step 8: Repeat steps 5, 6, and 7 to continuously cross-screen the subsets of the initial error model parent set until the final set of misaligned error models m is screened out. q , q is the number of cross-screening times; in each round of cross-screening, the error variables adjusted in the next round of cross-screening are different from those in the previous round of cross-screening, ensuring that each error variable is changed only once during the entire optical subsystem assembly and adjustment process; Step 9: The final set of offset error models m is selected q The misalignment error model best conforms to the changing law of the wavefront or star point when the optical subsystem is adjusted. It corresponds to the initial misalignment of the optical system. The adjustment of the optical subsystem can be completed by adjusting according to this initial misalignment.

2. A computer-aided alignment method based on cross-screening theory as claimed in claim 1, characterized in that: In step 1, if the auxiliary assembly complexity attribute is less than the preset complexity threshold, the optical system is not split; if the auxiliary assembly complexity attribute is greater than the preset complexity threshold, the optical system is split to reduce the computing pressure, and the value of the number of split groups ensures that the assembly complexity attribute of each group of subsystems obtained by splitting is less than the preset complexity threshold.

3. The computer-aided alignment method based on cross-screening theory according to claim 1, characterized in that: In step 4 and step 7, the parts with higher similarity are screened out by a preset screening ratio. Setting a higher screening ratio can ensure the accuracy of the desired result, while setting a lower screening ratio can speed up the convergence speed.

4. The computer-aided alignment method based on cross-screening theory according to claim 1, characterized in that: In step 9, according to the offset error model m q Interpolation calculations are performed to find a closer approximate solution to obtain a closer model of the offset error.

Images