Design method of mechanism-data combined driving high-integration uniform light and soft light microstructure element

Abstract

The invention discloses a mechanism-data combined driving high-integration uniform light and soft light microstructure element design method. The method comprises the following steps: firstly, constructing a mathematical model of a microstructure array through a functional modeling method; constructing a normal vector projection operator to realize superposition of a microstructure array and a macroscopic optical element; carrying out optical simulation analysis on the macroscopic optical element on which the microstructure array is superposed, and collecting light intensity index data; constructing and training a data-driven proxy model of the illumination distribution characteristics; the input of the model is parameters for defining the microstructure, and the output is a light intensity index; and in combination with a multi-objective optimization algorithm, performing Pareto frontier search in the trained data-driven agent model to obtain a microstructure parameter Pareto solution set meeting illumination uniformity and softness, and realizing integrated design of the combined optical element of the microstructure array and the macroscopic lens. The optical field regulation and control function of a traditional multi-stage optical system can be achieved through a single device, and the excellent effects of light beam shaping uniformity and softness are considered.

Description

Technical Field

[0001] The present invention belongs to the field of optical microstructure element design, and specifically relates to a mechanism-data jointly driven design method for highly integrated light homogenizing and softening microstructure elements. Background Art

[0002] With the rapid development of optical technology, optical microstructure elements are increasingly widely used in fields such as lighting, imaging, and optical communication. Especially in modern high-precision optical systems, higher requirements are put forward for the uniformity and softness of light sources. For example, in a projection display system, uneven light intensity distribution will lead to deterioration of imaging quality; in medical lighting equipment, too hard light sources will cause visual fatigue; in the field of laser processing, the non-uniformity of laser beams will seriously affect the processing accuracy and efficiency. Traditional beam shaping methods rely on macroscopic optical elements such as diffuser sheets, integrating rods, or complex lens groups. Although these methods can improve the light intensity distribution to a certain extent, they occupy too much space and are difficult to meet the urgent needs of modern optical systems for miniaturization and integration.

[0003] In recent years, optical microstructure elements based on micro-nano manufacturing technology have gradually become a research hotspot due to their advantages of high efficiency, compactness, and flexible design. By adding specific micro-nano structures on the material surface, precise control of the light field can be achieved, and thus the effects of light homogenization and softening can be achieved. However, existing microstructure design methods still face many challenges and deficiencies in practical applications. Many microstructure designs rely on empirical formulas or simple parametric models, and it is difficult to achieve stable control when dealing with complex light field distributions. Especially when high uniformity and softness requirements need to be met simultaneously, the ideal effect is often not achieved. Summary of the Invention

[0004] Aiming at the deficiencies of the existing technology, the present invention proposes a mechanism-data jointly driven design method for highly integrated light homogenizing and softening microstructure elements, and the specific technical solutions are as follows:

[0005] A mechanism-data jointly driven design method for highly integrated light homogenizing and softening microstructure elements, the method comprising the following steps:

[0006] Step 1: Construct a mathematical model of the microstructure array through a functional modeling method;

[0007] Step 2: Construct a normal vector projection operator, and based on the normal vector projection operator and the mathematical model of the microstructure array, superimpose the microstructure array on a macroscopic optical element to obtain a point cloud data file of the macroscopic optical element with the superimposed microstructure array;

[0008] Step 3: Conduct optical simulation analysis on the macroscopic optical element of the superposed microstructure array, and collect light intensity index data; construct and train a data-driven proxy model for the light distribution characteristics; the input of the data-driven model is the parameters defining the microstructure, and the output is the light intensity index; combined with a multi-objective optimization algorithm, perform a Pareto front search in the trained data-driven proxy model to obtain a Pareto solution set of microstructure parameters that meet the requirements of light uniformity and softness; according to the actual application requirements, select appropriate microstructure parameters from the Pareto solution set of microstructure parameters, and substitute them into Step 1 and Step 2 to obtain the point cloud data file of the macroscopic optical element with the optimal superposed microstructure array.

[0009] Further, the specific steps of Step 1 are as follows:

[0010] S1.1: Select and construct the array function of the microstructure array;

[0011] S1.2: Design a microstructure height function that defines the morphology of a single microstructure. The microstructure height function includes a basis function and a scaling factor function. The basis function is used to describe the contour morphology characteristics of a single microstructure in the array unit, and the scaling factor function is used to describe the overall variation law of the microstructure height in the global coordinate system.

[0012] Further, the construction of the microstructure array is specifically as follows: Establish a global coordinate system with the center of the entire microstructure array as the origin, and obtain the mapping from the global coordinate system to the local coordinate system with the center of each microstructure unit as the origin through the array function.

[0013] Further, the microstructure height function is the product of the basis function and the scaling factor function.

[0014] Further, the basis function is a polynomial function, and the scaling factor function is a Gaussian function.

[0015] Further, in Step 1, there are two microstructure arrays. The second microstructure array is interspersed in the gaps between the array units of the first microstructure array; the mathematical models of the two microstructure arrays are constructed using the same method, but the parameter values of the two microstructure arrays are different; the domain parameters of the second microstructure array satisfy the following conditions:

[0016]

[0017] Among them, R x , max , max is the domain that restricts the microstructures in each array unit of the first microstructure array, and R′ max is the domain that restricts the microstructures in each array unit of the second microstructure array, and d x is the distance between the center points of adjacent array units of the first microstructure array in the x coordinate direction, dy is the distance between the center points of adjacent array units of the first micro-structure array in the y coordinate direction.

[0018] Further, the second step includes the following sub-steps:

[0019] S2.1: Select the form of the superposition surface according to requirements, take the superposition surface as the base surface, and calculate the unit outer normal function g of the base surface;

[0020] S2.2: Multiply the micro-structure height function h by the unit outer normal function g of the base surface to obtain the normal projection operator N;

[0021] S2.3: Construct a computational grid, and the computational grid includes the x coordinates and y coordinates of the coordinate points required in the design or production and processing process;

[0022] S2.4: Combine the normal projection operator N and the computational grid point coordinates to obtain multiple sets of non-linear equations; each grid point corresponds to a set of non-linear equations;

[0023] S2.5: Use a high-precision parallel iterative algorithm to solve each set of non-linear equations respectively, solve the z coordinates of each grid point, and finally obtain the point cloud data file of the macro-optical element of the superposition micro-structure array.

[0024] Further, the high-precision parallel iterative algorithm is a hybrid method of the trust region-dogleg method and the Chebyshev iterative method. First, use the trust region-dogleg method to find a numerical solution with high confidence, and then use the Chebyshev iterative method to improve the numerical calculation accuracy.

[0025] Further, the third step specifically includes the following sub-steps:

[0026] S3.1: Construct multiple sets of parameter data of different micro-structure arrays, substitute the parameter data of the micro-structure array into the first step and the second step to obtain the point cloud data file of the macro-optical element of the superposition micro-structure array with different parameters; use optical simulation software to model the optical system obtained from the superposition micro-structure array and the macro-optical lens in the second step to obtain a coaxial optical system model; for the coaxial optical system model, after replacing the point cloud data file with different parameters, perform ray tracing simulation, collect the data of the illumination uniformity indexes L1, L2 and the softness index M of the optical system under different micro-structure parameters, establish a training data set, and use the Gaussian process regression algorithm to establish and train a data-driven proxy model of the illumination distribution characteristics; the input of the data-driven proxy model is the parameter of the micro-structure array defined in the first step, and the output is the illumination uniformity indexes L1, L2 and the softness index M of the coaxial optical system model corresponding to the micro-structure parameters;

[0027] S3.2: Utilize a multi-objective optimization algorithm, taking the illumination uniformity indicators L1 and L2 and the softness indicator M as optimization targets, and use the trained data-driven surrogate model to perform parameter optimization. Through iterative analysis, obtain the Pareto solution set of microstructure parameters that meets the illumination uniformity and softness requirements;

[0028] S3.3: Based on actual application requirements, select appropriate microstructure parameters from the Pareto solution set and bring them into steps 1 and 2 to obtain a point cloud data file of the macro-optical element superimposed with the optimal microstructure array.

[0029] Furthermore, the illumination uniformity evaluation index L1 is defined as the length of the illumination uniformity zone, which is the area from the edge sampling point where the illumination intensity increases to 95% of the central light intensity;

[0030] The light intensity uniformity evaluation index L2 is defined as: the root mean square of the relative light intensity of the sampling points in the uniform illumination area relative to the central sampling point;

[0031] The softness evaluation index M is defined as the average light intensity gradient at the edge of the light spot, that is, the average light intensity gradient in the transition zone of the light spot edge. The transition zone of the light spot edge is the area where the light intensity increases from 0 to 50% of the central light intensity.

[0032] The beneficial effects of the present invention are as follows:

[0033] The optical element designed by the highly integrated uniform light and soft light microstructure element design method of the present invention can realize the light field control function of the traditional multi-stage optical system through a single device, taking into account the excellent effects of beam shaping uniformity and softness, and meeting the needs of modern optical systems for miniaturization and integration. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 This is an overall implementation flow chart of an embodiment of the present invention.

[0035] Figure 2 The different h involved in the embodiments of the present invention c Microstructure basis function profile morphology under parameters.

[0036] Figure 3 This is the three-dimensional morphology of a periodic single microstructure involved in an embodiment of the present invention, where the left figure is a projection view of the periodic single microstructure, and the right figure is a three-dimensional view of the microstructure.

[0037] Figure 4 This is the three-dimensional morphology of the periodic composite microstructure involved in an embodiment of the present invention, where the left figure is a projection view of the microstructure and the right figure is a three-dimensional view of the microstructure.

[0038] Figure 5Schematic diagram of the microstructure involved in the embodiment of the present invention superimposed on the base surface in the normal projection manner.

[0039] Figure 6 Schematic diagram of the definition method for the light intensity uniform region and the edge transition region involved in the embodiment of the present invention.

[0040] Figure 7 Two-dimensional layout diagram of the optical model involved in the embodiment of the present invention.

[0041] Figure 8 Schematic diagram of the initial light intensity distribution of the Gaussian beam involved in the present invention.

[0042] Figure 9 Schematic diagram of the light intensity distribution of the Gaussian light after passing through the lens without microstructure involved in the present invention.

[0043] Figure 10 Schematic diagram of the light intensity distribution of the Gaussian light after passing through the microstructure element involved in the present invention. Detailed implementation manners

[0044] The present invention will be described in detail below according to the drawings and preferred embodiments. The purpose and effects of the present invention will become more apparent. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0045] The design method of the high-integration light homogenization and softening microstructure element driven by mechanism-data of the present invention designs the periodically arranged microstructure in a functional form and superimposes the microstructure array on the optical element in the normal projection manner to realize the integrated design of the combined optical element of the microstructure array and the macroscopic optical element. The specific implementation process is as Figure 1 shown.

[0046] I. Functional design of microstructure

[0047] By designing the array function, the arrangement mode of the microstructure is defined; by designing the microstructure height function, the morphology of a single microstructure is defined, and the single microstructures are neatly arranged in each array unit to complete the microstructure design. The following elaborates on each step in detail.

[0048] S1.1: Select and construct the array function of the microstructure array.

[0049] Construct the microstructure array form in various forms such as rectangular array and circular array. From the perspective of mathematical transformation, the array construction is actually a translation transformation operation on the original global coordinate system to obtain the local coordinate system with the center of each array unit as the coordinate origin. Therefore, first establish the global coordinate system with the center of the entire microstructure array as the origin, and then obtain the mapping from the global coordinate system to the local coordinate system with the center of each microstructure unit as the origin through the array function A.

[0050] In this embodiment, a square array is selected as the array pattern, and the array function expression is shown in the following formula (1):

[0051]

[0052] where X and Y are the local coordinates of each array element obtained after array transformation, x and y are the global coordinates before array transformation, and d x is the distance between the centers of adjacent array elements in the x - coordinate direction, and d y is the distance between the centers of adjacent array elements in the y - coordinate direction, satisfying d x = d y , is the floor symbol.

[0053] S1.2: Design a microstructure height function that defines the topography of a single microstructure. The microstructure height function includes a basis function and a scaling factor function. The expression of the microstructure height function h is shown in the following formula (2):

[0054] h = α·f(2)

[0055] where f is a reference function that describes the contour topography characteristics of a single microstructure in the array element, and is a composite function of the function A related to the microstructure array; α is a scaling factor function, which is used to describe the overall variation law of the microstructure height in the global coordinates and is a global control function of the microstructure height. In practical applications, the basis function f and the scaling factor function α can be in various forms such as polynomial functions, exponential functions, and trigonometric functions. In this embodiment, the basis function is selected as a fifth - degree polynomial function, and the expression is shown in the following formula (3):

[0056]

[0057] where R represents the distance from the points in each array element to the z - axis of the local coordinate system of the corresponding array, and is a composite function related to the array function A. The calculation method is shown in the following formula (4):

[0058]

[0059] R max limits the domain of the microstructure in each array element, satisfying R ≤ R max . To ensure that the microstructures do not interfere and overlap with each other, it is also necessary to ensure that R max is less than or equal to the array interval d. In this embodiment, R max = d, that is, the array microstructures just contact each other closely and are arranged tightly.

[0060] B, C, D, E, F, G are polynomial coefficients. To meet the requirement of microstructure smoothness, the following constraints are satisfied among the coefficients in this embodiment:

[0061]

[0062] Among them, h c is a parameter for controlling the profile of the basis function. The profiles of the basis functions under different h c parameters are as shown Figure 2 .

[0063] In this embodiment, the scaling factor function α is selected in the form of a Gaussian function, and the expression is as shown in the following formula (6):

[0064]

[0065] Among them, r max restricts the domain of definition of the microstructure in each array unit, satisfying r ≤ r max , where r represents the distance from the point in each array unit to the z-axis of the global coordinate system, and the calculation method is as shown in the following formula (7):

[0066]

[0067] Among them, ɑ c is the scaling factor value at the center of the domain of definition of the microstructure, that is, at r = 0, ɑ = ɑ c ; ɑ b is the scaling factor value at the boundary of the domain of definition of the microstructure, that is, at r = r max , α = α b . If α c = α b , then the scaling factor α is a constant. At this time, the microstructures in each array unit are scaled proportionally, and the profiles are exactly the same, showing a standard periodic arrangement. When α is a constant, the three-dimensional profile of the periodic microstructure is as shown Figure 3 .

[0068] According to steps S1.1 and S1.2 for microstructure design, only two parameters d x and d y in the array function, the parameter h c in the basis function, and the parameters r max , α c and α b in the scaling factor function need to be determined to completely define a set of periodically arranged microstructures. The height information of the periodically arranged microstructures is completely described by the microstructure height function h. In actual calculations, only the x coordinate and y coordinate of any coordinate point need to be given, and the microstructure height value corresponding to the point can be accurately calculated through the function h.

[0069] Furthermore, for the already arranged periodic microstructures, another set of periodic microstructures can be interspersed in the gaps between adjacent microstructure units in the same manner as in steps S1.1 and S1.2 to form a composite microstructure array. The three-dimensional topography of the periodic composite microstructure is as shown in Figure 4 shown, and the microstructures in this embodiment are in the form of composite microstructures.

[0070] The array function of the interspersed microstructures will be different, and the constant term in Equation (1) needs to be slightly adjusted. The basis function and scaling factor function of the interspersed microstructures can be the same as those in Equations (3) and (6), but different values of R max , h c , α c and α b parameters can be selected. To distinguish from the previously arranged microstructures, the defined parameters of the interspersed microstructures are denoted as R′ max , h′ c , α′ c and α′ b . It should be particularly noted that to avoid overlapping interference between the interspersed microstructures and the previously arranged microstructures, the domain R′ max of the interspersed microstructures needs to satisfy the following conditions:

[0071]

[0072] wherein, R max is the domain of the previously arranged microstructures, R′ max is the domain of the interspersed microstructures, and min is the minimum value function.

[0073] II. Calculation of the Normal Projection of the Microstructure

[0074] Construct a normal vector projection operator, and based on the normal vector projection operator and the mathematical model of the microstructure array, realize the superposition of the microstructure array and the macroscopic optical lens, specifically including:

[0075] S2.1: Select the form of the superposition surface (referred to as the base surface) according to requirements, and calculate the unit outer normal function g of the base surface.

[0076] In this embodiment, the form of the base surface is selected as the standard quadratic aspherical function, and the expression is as shown in Equation (9) below:

[0077]

[0078] where K is the conic coefficient and C is the paraxial curvature of the surface, i.e., the curvature at the vertex of the surface.

[0079] Through derivative operation, the unit outer normal function g of the base surface is:

[0080]

[0081] Among them, l, m, and n are the components of the function g in the x-axis, y-axis, and z-axis directions, respectively.

[0082] S2.2: Multiply the microstructure height function h and the unit outer normal function g of the base surface to obtain the normal projection operator N.

[0083] S2.3: Construct a computational grid. The computational grid includes the x-coordinates and y-coordinates of the coordinate points required in the design or production and processing processes. The z-coordinate of the coordinate points needs to be obtained by solving equations. The grid form can be various forms such as a rectangular coordinate grid and a cylindrical coordinate grid.

[0086] Figure 5 S2.4: Combine the normal projection operator N and the computational grid point coordinates to obtain a set of nonlinear equations, as shown in the following formula (11):

[0087]

[0088] One grid point corresponds to a set of nonlinear equations. x0, y0, and z0 are the coordinates of the base points, and x, y, and z are the coordinates of the base points after superimposing the microstructure height in the normal direction. The process of superimposing the microstructure on the base surface in the normal projection manner is as

[0089] shown.

[0090] S2.5: Use a high-precision parallel iterative algorithm to solve each set of nonlinear equations respectively. Solve the x0 and y0 coordinates through the first two equations in the equations, and substitute them into the third equation to obtain the z-coordinates of each computational base point.

[0091] In this embodiment, to achieve parallel computing, use the scientific computing software MATLAB for programming. Store the x-coordinates and y-coordinates of each computational grid point in the form of matrix variables respectively, and achieve parallel computing through the operation method of matrix dot multiplication. Solve iteratively through the iterative algorithm. Finally, obtain the z-coordinate matrix of the computational grid points that meets the accuracy requirements, and finally output a point cloud file containing microstructure information. Among them, the iterative algorithm selects a hybrid method of the trust region-dogleg method and the Chebyshev iterative method. First, use the trust region-dogleg method with high stability to find a numerical solution with high confidence, and then use the Chebyshev iterative method to improve the numerical calculation accuracy.

[0092] III. Data-driven optimization of microstructure parameters

[0093] Optical simulation analysis is carried out on the macroscopic optical element of the superimposed microstructure array, and the light intensity index data is collected; a data-driven proxy model of the light distribution characteristics is constructed and trained; the input of the data-driven model is the parameters defining the microstructure, and the output is the light intensity index; combined with the multi-objective optimization algorithm, Pareto front search is carried out in the trained data-driven proxy model to obtain the Pareto solution set of the microstructure parameters that meet the light uniformity and softness; according to the actual application requirements, appropriate microstructure parameters are selected from the Pareto solution set of the microstructure parameters, and substituted into Step 1 and Step 2 to obtain the point cloud data file of the macroscopic optical element with the optimal superimposed microstructure array. Specifically, it includes the following sub-steps:

[0091] S3.1: Construct multiple groups of parameter data of different microstructure arrays through Latin hypercube sampling or other construction methods, and substitute the parameter data of the microstructure array into Step 1 and Step 2 to obtain the point cloud data files of the macroscopic optical elements of the superimposed microstructure arrays with different parameters; use optical simulation software to model the optical system obtained by superimposing the microstructure array and the macroscopic optical lens in Step 2 to obtain a coaxial optical system model; for the coaxial optical system model, after replacing the point cloud data files with different parameters, perform ray tracing simulation, collect the light uniformity indexes L1, L2 and the softness index M of the optical system under different microstructure parameters, establish a training data set, and establish and train a data-driven proxy model of the light distribution characteristics using the Gaussian process regression algorithm; the input of the data-driven proxy model is the parameters of the microstructure array defined in Step 1, and the output is the light uniformity indexes L1, L2 and the softness index M of the coaxial optical system model corresponding to the microstructure parameters.

[0092] The light uniformity evaluation index L1 is used to quantify the size of the light-uniform area, L2 is used to quantify the intensity distribution consistency of the light field in the target area, and the softness index M quantifies the edge transition characteristics of the light field.

[0093] In this embodiment, the light uniformity index L1 is defined as: starting from the edge sampling point, when the light intensity increases to 95% of the central light intensity, it enters the light-uniform area, and the size of the uniform area is measured by the length L1, as Figure 6 shown.

[0094] The light intensity uniformity of the light-uniform area is characterized by L2, which is the root mean square of the relative light intensity of the sampling points with respect to the central sampling point:

[0095]

[0096] where n is the number of light intensity sampling points, is the relative light intensity of the sampling point, defined as the ratio of the sampling point light intensity to the central sampling point light intensity, is the average value of the relative light intensity. The smaller the value of L2, the higher the light intensity uniformity.

[0097] The light softness index M is defined by the average light intensity gradient at the edge of the light spot. Starting from the edge sampling points, the area where the light intensity increases from 0 to 50% of the central light intensity is defined as the light spot edge transition region, as Figure 6 shown. The average light intensity gradient in the edge transition region is defined by the following formula (13):

[0098]

[0099] where w is the width of the light spot edge transition region, and I r＝0.5 is the actual light intensity corresponding to where the relative light intensity is 0.5.

[0100] In this embodiment, the inputs of the surrogate model are the parameters R max , h c , α c , α b , R′ max , h′ c , α′ c and α′ b , and the outputs are the light intensity uniformity indexes L1, L2 and the softness index M of the coaxial optical system model under the corresponding microstructure parameters.

[0101] S3.2: Using a multi-objective optimization algorithm, with the light intensity uniformity indexes L1, L2 and the softness index M as the optimization objectives, and using the trained data-driven surrogate model for parameter optimization. Through iterative analysis, a Pareto solution set of microstructure parameters that meet the requirements of light intensity uniformity and softness is obtained. The Pareto solution set contains multiple groups of different microstructure definition parameters.

[0102] S3.3: According to the actual application requirements, select appropriate microstructure parameters from the Pareto solution set and substitute them into Step 1 and Step 2 to obtain the point cloud data file of the macro-optical element with the superimposed optimal microstructure array.

[0103] In this embodiment, the multi-objective optimization algorithm selects a hybrid algorithm of the NSGA-II algorithm and the goal-attain method. First, use the NSGA-II algorithm to find the Pareto solution set with low confidence, and then use the goal-attain method to refine the optimization of the solution set to obtain the final Pareto solution set with high confidence. Select the solution in the solution set that minimizes the light intensity uniformity index L2 of the optical system as the optimal solution. Figure 9 and Figure 10They are the light intensity distributions of Gaussian light passing through the lens without microstructure and the microstructure element with the optimal solution parameters, respectively. It can be found that after adding the microstructure, the light intensity uniformity is significantly improved. The width L of the light intensity uniform region increases from 30 mm to 49 mm, the light intensity uniformity index L2 decreases from 0.0148 to 0.0089, and the average light intensity gradient in the transition region at the edge of the light spot also decreases across magnitudes from 1.03×10 -5 w / mm 3 to 1.18×10 -6 w / mm 3 , and the softness of the light intensity is significantly improved.

[0104] Those of ordinary skill in the art can understand that the above are only preferred examples of the invention and are not used to limit the invention. Although the invention has been described in detail with reference to the foregoing examples, for those skilled in the art, they can still modify the technical solutions described in the foregoing examples, or perform equivalent replacements for some of the technical features. Any modifications, equivalent replacements, etc. made within the spirit and principle of the invention shall be included within the protection scope of the invention.

Claims

1. A design method for a highly integrated light homogenizing and softening microstructure element driven by the combination of mechanism and data, characterized in that The method includes the following steps: Step 1: Construct a mathematical model of the microstructure array by means of functional modeling method; Step 2: Construct a normal vector projection operator, and based on the normal vector projection operator and the mathematical model of the microstructure array, superpose the microstructure array on the macroscopic optical element to obtain a point cloud data file of the macroscopic optical element with the superposed microstructure array; Step 3: Conduct optical simulation analysis on the macroscopic optical element with the superposed microstructure array, and collect light intensity index data; construct and train a data-driven proxy model for the light distribution characteristics; the input of the data-driven model is the parameters defining the microstructure, and the output is the light intensity index; combined with a multi-objective optimization algorithm, conduct Pareto front search in the trained data-driven proxy model to obtain a Pareto solution set of microstructure parameters that meet the requirements of light uniformity and softness; according to the actual application requirements, select appropriate microstructure parameters from the Pareto solution set of microstructure parameters, substitute them into Step 1 and Step 2, and obtain a point cloud data file of the macroscopic optical element with the optimal superposed microstructure array.

2. The mechanism-data jointly driven high-integration light homogenizing and softening microstructure element design method according to claim 1, characterized in that The specific steps of Step 1 include the following sub-steps: S1.1: Select and construct an array function of the microstructure array; S1.2: Design a microstructure height function that defines the morphology of a single microstructure. The microstructure height function includes a basis function and a scaling factor function. The basis function is used to describe the contour morphology characteristics of a single microstructure in the array unit, and the scaling factor function is used to describe the overall change law of the microstructure height in the global coordinate system.

3. The design method of a highly integrated light homogenizing and softening microstructure element driven by mechanism-data joint, according to claim 2, is characterized in that The construction of the microstructure array is specifically as follows: Establish a global coordinate system with the center of the entire microstructure array as the origin, and obtain the mapping from the global coordinate system to the local coordinate system with the center of each microstructure unit as the origin through the array function.

4. The mechanism-data jointly driven high-integration light homogenizing and softening microstructure element design method according to claim 2, characterized in that, The microstructure height function is the product of the basis function and the scaling factor function.

5. The mechanism-data jointly-driven high-integration light homogenizing and softening microstructure element design method according to claim 4, characterized in that The basis function is a polynomial function, and the scaling factor function is a Gaussian function.

6. The design method of a highly integrated light homogenization and softening microstructure element driven by mechanism-data joint is characterized in that In Step 1, there are two microstructure arrays. The second microstructure array is interspersed in the gaps between the array units of the first microstructure array; the mathematical models of the two microstructure arrays are constructed by the same method, but the parameter values of the two microstructure arrays are different; the domain parameters of the second microstructure array satisfy the following conditions: Among them, R max is the domain that restricts the microstructures in each array unit of the first micro-structure array, and R ′ max is the domain that restricts the microstructures in each array unit of the second micro-structure array. d x is the distance between the center points of adjacent array units of the first micro-structure array in the x-coordinate direction, and d y is the distance between the center points of adjacent array units of the first micro-structure array in the y-coordinate direction.

7. The design method of the high-integration uniform light and soft light microstructure element driven by mechanism-data joint is characterized in that, Step 2 includes the following sub-steps: S2.1: Select the form of the superposition surface according to the requirements, take the superposition surface as the base surface, and calculate the unit outer normal function g of the base surface; S2.2: Multiply the microstructure height function h and the unit outer normal function g of the base surface to obtain the normal vector projection operator N; S2.3: Construct a computational grid, and the computational grid includes the x coordinates and y coordinates of the coordinate points required in the design or production and processing process; S2.4: Combine the normal vector projection operator N and the computational grid point coordinates to obtain multiple groups of nonlinear equations; each grid point corresponds to a group of nonlinear equations; S2.5: Use a high-precision parallel iterative algorithm to solve each group of nonlinear equations respectively, solve the z coordinates of each grid point, and finally obtain a point cloud data file of the macroscopic optical element with the superposed microstructure array.

8. The design method of the high-integration uniform light and soft light microstructure element driven by mechanism-data joint according to claim 7, characterized in that The high-precision parallel iterative algorithm is a hybrid method of the trust region-dogleg method and the Chebyshev iterative method. First, the trust region-dogleg method is used to find a numerical solution with high confidence, and then the Chebyshev iterative method is used to improve the numerical calculation accuracy.

9. The design method of the high-integration light homogenizing and softening microstructure element driven by mechanism-data joint driving according to claim 1, characterized in that The specific steps of step three include the following sub-steps: S3.1: Construct multiple groups of parameter data of different micro-structure arrays, and substitute the parameter data of the micro-structure arrays into step one and step two to obtain point cloud data files of the macroscopic optical elements of the superimposed micro-structure arrays with different parameters; Use optical simulation software to model the optical system obtained by superimposing the micro-structure array and the macroscopic optical lens in step two to obtain a coaxial optical system model; for the coaxial optical system model, after replacing the point cloud data files with different parameters, perform ray tracing simulation, collect the data of the illumination uniformity indexes L1, L2 and the softness index M of the optical system under different micro-structure parameters, establish a training data set, and use the Gaussian process regression algorithm to establish and train a data-driven proxy model of the illumination distribution characteristics; the input of the data-driven proxy model is the parameter of the micro-structure array defined in step one, and the output is the illumination uniformity indexes L1, L2 and the softness index M of the coaxial optical system model under the corresponding micro-structure parameters; S3.2: Use the multi-objective optimization algorithm, with the illumination uniformity indexes L1, L2 and the softness index M as the optimization objectives, and use the trained data-driven proxy model to perform parameter optimization. Through iterative analysis, obtain the Pareto solution set of the micro-structure parameters that meet the requirements of illumination uniformity and softness; S3.3: According to the actual application requirements, select appropriate micro-structure parameters from the Pareto solution set, and substitute them into step one and step two to obtain the point cloud data file of the macroscopic optical element of the superimposed optimal micro-structure array.

10. The design method of the high-integration light homogenization and softening microstructure element driven by mechanism-data joint, according to claim 9, is characterized in that, The illumination uniformity evaluation index L1 is defined as the length of the illumination uniform area, and the illumination uniform area is the area starting from the edge sampling point where the illumination intensity increases to 95% of the central light intensity; The light intensity uniformity evaluation index L2 is defined as: the root mean square of the relative light intensity of the sampling points in the illumination uniform area relative to the central sampling point; The softness evaluation index M is defined as the average light intensity gradient at the edge of the light spot, that is, the average light intensity gradient in the light spot edge transition area, and the light spot edge transition area is the area where the illumination intensity increases from 0 to 50% of the central light intensity.

Images