TWI664545B - Method for producing a primary optical free-form surface structure - Google Patents

Abstract

The invention discloses a manufacturing method of a primary optical free-form surface structure, which includes obtaining an image corresponding to a target light shape. Find the preliminary objective function. Obtain a preliminary free-form solid model. Import the beam tracing software to simulate the preliminary result light shape. Find the preliminary result function corresponding to the preliminary result light shape. According to the comparison result between the preliminary result function and the preliminary objective function, a modified objective function is generated. Obtain a modified freeform solid model. The beam tracing software is introduced to simulate the corrected light shape. Find the correction result function. The modified result function and the preliminary objective function are compared. When the difference between the two is lower than a preset threshold value, the primary optical free-form surface structure is made according to the modified free-form surface solid model.

Description

Manufacturing method of primary optical free-form surface structure

The invention relates to a method for manufacturing an optical structure, in particular to a method for manufacturing a primary optical free-form surface structure.

Before a light-emitting diode (LED for short) is actually used as a component of a lighting device, the optical structure design is generally performed twice. First, when LEDs are made into the required optoelectronic components through integrated circuit (IC) packaging technology, the first optical structure design must be performed first to solve the LED light angle, light intensity, luminous flux, Light intensity distribution and the range and distribution of color temperature. Generally, the light emitted by high-power LEDs passes through the primary optical structure lens, and its light emitting angle is about 120 degrees. Based on this, the secondary optical structure design is performed. The purpose is to re-emit the light emitted by the primary optical structure lens. Further adjust its light angle and change its optical performance.

Under the current design structure of common LED structures, to adjust the light emitted by the LED to form a special light shape, a free-form surface design is usually performed on the surface of the secondary optical structure. The micro-processing problems of free-form lenses and the manufacturing cost of lenses have limited the application of free-form optical lenses in the market. Existing algorithms for free-form surface optics mainly include Tailored, Partial Differential Equation (PDE), Simultaneous Multiple Surface (SMS), and geometric algorithms.

Among them, the main concept of the trimming method is to establish a nonlinear partial differential equation system on the optical surface shape according to the illumination distribution of the target surface and the characteristics of the light source. The nonlinear surface differential equation system is obtained by solving this nonlinear partial differential equation system. However, this One method is less suitable for extended light sources. Similarly, the PDE method is also less suitable for extended light sources. The extended light source refers to a light source with a relatively large light emitting area (compared to a point light source). Since the size of the light source is relatively large, there are many directions of light emission. Existing light sources used in LED packages are generally extended light sources (body light sources rather than point light sources) compared to their packages.

Furthermore, the main concept of the SMS method is to first establish the correspondence between two pairs of input wavefronts and two pairs of output wavefronts according to the light energy distribution characteristics of the light source and the target plane, and then design two free-form surfaces of an optical system at the same time. So that the incident wavefront can be in one-to-one correspondence with the outgoing wavefront after being refracted or reflected by the two free-form surfaces. However, its disadvantage is that when the light spot is diffuse, it cannot guarantee that the light in the middle part can be uniformly distributed in the corresponding area on the target plane.

On the other hand, the main concept of the geometric algorithm is to establish the relationship between the incident angle and the free-form surface based on the non-imaging lighting theory and the geometric relationship between the free-form surface and the light, so as to obtain the free-form surface.

As mentioned above, the current common LED structure design architectures are mainly free-form surface designs for the surface of the secondary optical structure. The main reason is that they are limited by micro-processing issues and lens manufacturing cost issues. However, this method not only complicates the overall structure and is not conducive to miniaturization of the device, but also because light must be converted through more layers of medium before reaching the surface of the target illuminated object, which will cause a large energy loss. Therefore, how to solve the aforementioned problems and complete the required free-form surface structure design on a primary optical structure has become one of the important issues to be solved urgently.

The technical problem to be solved by the present invention is to provide a manufacturing method of a primary optical free-form surface structure for the shortcomings of the prior art. The objective function and the correction result function calculated by simulation are used to find a difference between the two, and repeatedly adjust the modified objective function when the difference value is not lower than a preset threshold value, thereby obtaining the best modified free-form surface volume. Model, and make a primary optical free-form surface structure that can produce the desired target light shape.

In order to solve the above technical problems, one of the technical solutions adopted by the present invention is to provide a method for manufacturing a primary optical free-form surface structure, which includes: (A) obtaining an image corresponding to a target light shape; (B) Analyze the image to obtain a preliminary objective function corresponding to the target light shape; (C) perform a free-form surface operation of the point light source according to the preliminary objective function to obtain a preliminary free-surface solid model; (D) ) Importing the preliminary free-form surface solid model into a beam tracing software to simulate a preliminary result light shape; (E) analyzing the preliminary result light shape to obtain a preliminary result function corresponding to the preliminary result light shape (F) comparing the preliminary result function and the preliminary objective function; (G) generating a modified objective function according to a comparison result of the preliminary result function and the preliminary objective function; (H) according to the modification The objective function performs a free-form surface operation of the volume light source to obtain a modified free-form solid model; (I) imports the modified free-form solid model into the beam tracking software to simulate a Positive result light shape; (J) analyzing the correction result light shape to obtain a correction result function corresponding to the correction result light shape; (K) comparing the correction result function and the preliminary objective function to obtain Obtain a difference between the two; (L) determine whether the difference is lower than a preset threshold; (M) when the difference is not lower than the preset threshold, repeat step (G) To (L); and (N) when the difference value is lower than the preset threshold value, according to the modified free-form solid model, to make the primary optical free-form structure.

One of the beneficial effects of the present invention is that the method for making a one-time optical free-form surface structure provided by the present invention can obtain a value between one of the two by “comparing a preset objective function with a correction result function calculated by simulation”. "Difference value" and "Adjust the correction objective function repeatedly when the difference value has not fallen below a preset threshold value" to obtain the best modified free-form solid model, and make it An optical free-form surface structure that produces the desired target light shape.

In order to further understand the features and technical contents of the present invention, please refer to the following detailed description and drawings of the present invention. However, the drawings provided are only for reference and description, and are not intended to limit the present invention.

P1, P2‧‧‧Image

F1‧‧‧ preliminary objective function

F2‧‧‧Result function

F3‧‧‧ modified objective function

L‧‧‧light trail

M‧‧‧Freeform solid model

FIG. 1 is a flowchart of a manufacturing method according to a first embodiment of the present invention.

FIG. 2 is a schematic diagram of an image corresponding to a target light shape in the first embodiment of the present invention.

FIG. 3 is a schematic diagram of a preliminary objective function corresponding to a target light shape in the first embodiment of the present invention.

FIG. 4 is a schematic diagram of a free-form three-dimensional model simulated according to a preliminary objective function in the first embodiment of the present invention.

FIG. 5 is a schematic diagram of simulating a light ray trajectory by importing a free-form three-dimensional model into a beam tracing software in the first embodiment of the present invention.

FIG. 6 is a schematic diagram of the resulting light shape simulated by the beam tracing software in the first embodiment of the present invention.

FIG. 7 is a schematic diagram of a correction result function corresponding to the resulting light shape in the first embodiment of the present invention.

FIG. 8 is a schematic diagram of further generating a modified objective function according to a comparison result between a result function and an objective function in the first embodiment of the present invention.

FIG. 9 is a flowchart of steps of generating the modified objective function according to the second embodiment of the present invention.

The following is a description of specific embodiments to describe the implementation of the “method for making a primary optical free-form surface structure” disclosed by the present invention. Those skilled in the art can understand the advantages and effects of the present invention from the contents disclosed in this specification. The present invention can be implemented or applied through other different specific embodiments, and the details in this specification can also be implemented based on different viewpoints and applications without departing from the concept of the present invention. Modifications and changes. In addition, the drawings of the present invention are merely a schematic illustration, and are not drawn according to actual dimensions, and are stated in advance. The following embodiments will further describe the related technical content of the present invention in detail, but the disclosed content is not intended to limit the protection scope of the present invention.

It should be understood that although the terms first, second, third, etc. may be used herein to describe various elements or signals, etc., these elements or signals should not be limited by these terms. These terms are used to distinguish one element from another, or a signal from another signal. In addition, as used herein, the term "or" may include all combinations of any one or more of the associated listed items, as the case may be.

[First embodiment]

Please refer to FIG. 1 to FIG. 8, FIG. 1 is a flowchart of a manufacturing method of a first embodiment of the present invention; FIG. 2 is a schematic diagram of an image corresponding to a target light shape in the first embodiment of the present invention; A schematic diagram of the preliminary objective function corresponding to the target light shape in the first embodiment of the invention; FIG. 4 is a schematic diagram of a free-form three-dimensional model simulated according to the preliminary objective function in the first embodiment of the invention; FIG. 5 is a diagram of the first embodiment of the invention Importing a free-form three-dimensional model into a beam tracing software to simulate its light trajectory; Figure 6 is a schematic diagram of the resulting light shape simulated by the beam tracing software in the first embodiment of the present invention; Schematic diagram of the correction result function of the result light shape; FIG. 8 is a schematic diagram of further generating the correction objective function according to the comparison result between the result function and the objective function in the first embodiment of the present invention. As can be seen from the above figures, the first embodiment of the present invention provides a method for manufacturing a one-time optical free-form surface structure, which includes analyzing a target light shape to obtain a corresponding function (as shown in FIG. 2 and FIG. 3), and initially establishing a model ( (As shown in Fig. 4), the light shape simulated by the beam tracing software (as shown in Fig. 5), the result light shape is analyzed and the corresponding function (as shown in Figs. 6 and 7) is obtained, and the comparison result light shape and the target When the function obtained by the light shape (as shown in FIG. 8) and other steps are compared, and when the difference between the result of the light shape and the target light shape is less than a preset threshold value, according to the corresponding free-form solid model, Made into a light Learn freeform structure. In the following, with reference to the main flow shown in FIG. 1, the operation details of each step will be explained through the drawings of FIGS. 2 to 8.

First, please refer to FIG. 1, FIG. 2, and FIG. 3. In the preferred embodiment of the present invention, the first thing to be established is the target light shape expected to be generated by the LED. In terms of specific industrial applications, on the one hand, it can be the best lighting simulation graphics obtained by software for specific applications (such as bedrooms, living rooms, conference rooms, or stages); on the other hand, it can also be actual Apply the lighting results in a specific place (such as the original use of other light sources to generate lighting, you want to replace the LED with the same optical structure to provide the same lighting), and obtain the target lighting result graphics by shooting and other methods. In this embodiment, the present invention first obtains an image P1 corresponding to the target light shape (as shown in FIG. 2 and step S200 in FIG. 1), and analyzes the image P1 by an image processing program to obtain a preliminary corresponding light target shape. The objective function F1 (as shown in FIG. 3, step S202 in FIG. 1). As mentioned earlier, the image P1 may be an image file extracted from an analog frame, or an image frame file stored after actual shooting.

In this embodiment, the analysis method specifically performed is to obtain a corresponding preliminary objective function F1 according to the illuminance value of each point of the image P1. It is worth mentioning that, although for the convenience of explaining the main structure of the present invention, a flat graphic is used as the image P1 corresponding to the target light shape in this embodiment, but the application of the present invention is not used as such. Limitation, in addition to the previously mentioned simulation graphics can be in 3 dimensions (abbreviated as 3D) format, it can also be used for 3D optics with computer-aided verification (Computer Add Verification) The measurement technology first analyzes the surface of the object to be illuminated to obtain its 3D three-dimensional structure, and cooperates with multi-angle photography and other technologies to obtain the target illuminance of each point on the surface of the object. In short, the image P1 is not limited to a flat graphic.

On the other hand, when the corresponding preliminary objective function F1 is obtained based on the image P1, the calculation can be performed by the current common optical simulation software (such as LightTools, TracePro, ASAP, OSLO, or ZEMAX, but not limited to this). . The specific details and principles are not repeated here.

After obtaining the preliminary objective function F1, the present invention first assumes that the light source providing illumination is a point light source, and performs the most preliminary free-form surface operation. As we all know, for the structure of the package, the LED light source is not a true point light source, and can often be regarded as a light source that can emit light at 180 degrees and has Lambertian characteristics. Its light intensity distribution is centered Strongest but weaker in the periphery. However, in order to meet the limitations of the existing optical simulation software, the present invention performs a free-form surface calculation with a point light source, and obtains a preliminary free-form solid model M (as shown in FIG. 4 and step S204 in FIG. 1). The principle on which this part is based is mainly through the relationship between energy conservation and energy mapping plus Snell`s Law, which can be used to infer the required free-form surface. In the process of performing a free-form surface calculation to obtain a preliminary free-form solid model M, some preset conditions (such as the diameter or height of a lens) are set for the structure of the free-form solid model M, so that the produced product Able to meet the assembly needs in practical use.

If an actual primary optical free-form surface structure is made directly from the aforementioned preliminary free-form solid model M, the light shape projected at this time cannot meet the originally expected target light shape. The reason is as mentioned above, because the actual LED light source is not a point light source, there is a deviation from the ideal point light source, and the light intensity near the center will be higher than the surrounding light intensity. Therefore, the present invention does not directly use the optical free-form surface structure required for the actual production of the preliminary free-form three-dimensional model M directly, but rather introduces this preliminary free-form three-dimensional model M into the beam tracing software for simulation by the beam tracing software. The LED light source passes the light trajectory L projected by this preliminary free-form three-dimensional model M (as shown in FIG. 5), and according to the irradiation result of the light trajectory L projected on the surface of the target object, the corresponding light shape (such as (Step S206 of FIG. 1).

When simulating the light trajectory projected by the LED light source through the beam tracing software, the functions of the aforementioned software such as LightTools or TracePro can also be used (the same is not limited to this). Specifically, in this step, the effect of the selected material properties on the trajectory of the light is actually considered. On the related optical application software, the material name and interpolation can be further defined through the material editor or other similar functions. (interpolation), temperature, refractive index, or absorption wavelength. More importantly, in this step, the light-emitting characteristics of the LED light source are also actually considered for simulation. The technical details of this part will be discussed in detail later.

Please refer to FIG. 2 and FIG. 6 together. As mentioned earlier, since the light intensity of the actual LED light source near the center is higher than that of the surrounding light, if the result of the simulation result is The image P2 represents and compares the image P2 of FIG. 6 with the image P1 of FIG. 2. It can be found that the image P2 corresponding to the resulting light shape is not consistent with the image P1 of the originally desired target light shape. At this time, in order to understand the difference between the two through an objective comparison benchmark and further perform subsequent correction steps, the present invention analyzes the image P2 through an image processing program to obtain a preliminary result function F2 corresponding to the resulting light shape. (As shown in FIG. 7, step S208 in FIG. 1).

When comparing the preliminary objective function F1 corresponding to the target light shape with the preliminary result function F2 corresponding to the resulting light shape (as shown in FIG. 8 and step S210 in FIG. 1), it can be clearly found that the intermediate peak value of the preliminary result function F2 is relatively The intermediate peak value of the preliminary objective function F1 comes high. Near the periphery, the value of the preliminary result function F2 is lower than the value of the preliminary objective function F1. It can be seen that the preliminary objective function F1 is directly imported into the system, and the result function F2 obtained by executing steps S204 to S208 will not be consistent with the preliminary objective function F1. In other words, if the preliminary objective function F1 is directly imported into the system and executed, The free-form surface solid model M obtained in step S204 is used to produce an optical free-form surface structure, and the projected light shape will not be the ideal target light shape. However, the purpose of the present invention is to find a primary optical free-form surface structure capable of projecting an ideal target light shape. Therefore, it is necessary to find a free-form three-dimensional model M, and a result function F2 corresponding to the projected light shape. A sufficient degree of consistency can be achieved with the preliminary objective function F1 generated by analyzing the image P1 corresponding to the target light shape.

Specifically, the present invention compares the result function F2 with the preliminary objective function F1 at this time, and further generates a set of modified objective functions F3 according to the comparison result (see step S212 in FIG. 1). In this embodiment, the specific adjustment method is to maintain the preliminary objective function On the premise that the area under the curve of F1 is constant, the values of the coordinates of the preliminary objective function F1 are adjusted according to the preliminary result function F2 to obtain the values of the coordinates of the modified objective function F3. Specifically, when the value of the preliminary result function F2 at a specific coordinate position is higher than the value of the preliminary objective function F1 at the specific coordinate position, the value of the preliminary objective function F1 at the specific coordinate position is reduced. On the other hand, at a position where the value of the preliminary result function F2 is lower than the value of the preliminary objective function F1, the value of the preliminary objective function F1 is raised. Through the foregoing method, a modified objective function F3 is generated. It must be particularly mentioned here that the method for obtaining the modified objective function F3 in the present invention is not limited to this.

Next, the present invention uses the modified objective function F3 as an input function, and performs freeform surface calculation again to obtain a modified freeform solid model M corresponding to the modified objective function F3 (see step S214 in FIG. 1). Similarly, the modified free-form three-dimensional model M corresponding to the modified objective function F3 is then introduced into the beam tracing software to simulate the illumination result of the LED light source projected on the surface of the target object by the modified free-form three-dimensional model M through the beam tracing software. To simulate the light shape of the corresponding correction result (see step S216 in FIG. 1). After simulating the corresponding correction result light shape, the result function F2 corresponding to the correction result light shape is obtained according to the correction result light shape (as shown in step S218 in FIG. 1), so as to match the "preliminary objective function F1" (instead of the modified objective function). F3) perform an objective comparison between them, and find a difference between the two (see step S220 in FIG. 1).

As mentioned earlier, the present invention expects to obtain a result function F2 with a sufficient degree of consistency with the preliminary objective function F1. Therefore, the result function F2 corresponding to the light shape of the correction result is phased with the preliminary objective function F1. After comparing and obtaining the aforementioned difference value, the present invention compares the difference value with a preset threshold value set in advance (see step S222 in FIG. 1).

It is quite easy to understand that when the difference between the result function F2 and the preliminary objective function F1 is smaller, it indicates that the higher the similarity between the two, the more it meets the purpose of the present invention.

Therefore, if the difference between the result function F2 and the preliminary objective function F1 can be less than a preset threshold value, it means that the modified free-form solid model M obtained by performing the free-form operation through the modified objective function F3 is obtained by the beam tracing software. The simulated light shape and the target light shape are consistent with requirements, and therefore, it can proceed to step S224, and a primary optical free-form structure made according to the modified free-form solid model M.

Conversely, if the difference between the result function F2 and the preliminary objective function F1 is greater than a preset threshold value, it means that the modified free-form surface solid model M obtained by modifying the objective function F3 cannot yet produce a primary optical free-form surface that meets the requirements. structure. At this time, according to the comparison result of step S220, the preliminary objective function F1 must be modified again to obtain a new set of modified objective functions F3. Because the deviation between the actual LED light source and the ideal point light source has been corrected in the previous steps, compared with the previous result function F2 obtained directly through the preliminary objective function F1, the objective function F3 has been modified through a series of steps. The difference between the obtained function F2 and the preliminary objective function F1 becomes smaller. In the process of repeatedly performing steps S212 to S220, a smaller difference value is gradually obtained (that is, the result function F2 approaches the preliminary objective function F1), and when the difference value can be less than a preset threshold value, a match is obtained. The required modified free-form three-dimensional model M is performed, and a mold flow analysis is performed according to the modified free-form three-dimensional model M to establish a mold required for producing an optical free-form surface structure, and then the required optical free-form surface structure is made from the mold.

Based on the above description, with the process shown in FIG. 1 and the component symbols shown in FIG. 2 to FIG. 8, the main process structure of the present invention is initially arranged as follows: Step S200: obtaining an image P1 corresponding to the target light shape; Step S202: Analyze the image P1 to obtain a preliminary objective function F1 corresponding to the target light shape; step S204: perform a free-form surface operation of the point light source according to the preliminary objective function F1 to obtain a preliminary free-form solid model M; Step S206: import the preliminary free-form surface solid model M into the beam tracing software to simulate the preliminary result light shape; step S208: analyze the preliminary result light shape to obtain a preliminary result function F2 corresponding to the preliminary result light shape; step S210: Compare the preliminary result function F2 and the preliminary objective function F1; step S212: generate a modified objective function F3 according to the comparison result between the preliminary result function F2 and the preliminary objective function F1; step S214: perform a free-form surface operation according to the modified objective function F3 to obtain a correction Free-form three-dimensional model; Step S216: Import the modified free-form three-dimensional model into the beam tracking software to simulate the light shape of the correction result; Step S218: Analyze the light shape of the correction result to obtain the correction result function F2 corresponding to the light shape of the correction result; Step S220: Compare the correction result function F2 and the preliminary objective function F1 to obtain a difference value between the two; Step S222: determine whether the difference value is lower than a preset threshold value, and when the difference value is not lower than the preset threshold value , Repeat step S212 to step S222, otherwise, go to step S224; and step S224: according to the modified free song The three-dimensional surface model is made into an optical free-form surface structure.

[Second embodiment]

As mentioned above, when the difference value between the result function F2 and the preliminary objective function F1 is not less than (greater than or equal to) a preset threshold value, a set of modified objective functions F3 is further generated, and the present invention obtains the modified objective function F3. The method is not limited to the method described in the first embodiment.

Please refer to FIG. 1 and FIG. 9 for reference. FIG. 9 is a flowchart of steps for generating the modified objective function according to the second embodiment of the present invention. In a second embodiment of the invention, First, a difference function is obtained according to the comparison result between the result function F2 and the preliminary objective function F1, and a large number of modified objective functions F3 are generated randomly according to the aforementioned difference function (step S300), and according to step 214 shown in FIG. 1 From step S218, a correction result function F2 corresponding to each of the respective correction objective functions F3 is generated (step S302).

The main purpose of the aforementioned process is to first generate a large number of modified objective functions F3 through random number operations in the numerical interval between the result function F2 and the preliminary objective function F1, and based on this, establish many sets of modified objective functions F3 and modifications. The result function F2. The main difference between it and the first embodiment is that the first embodiment is to re-do the preliminary objective function F1 with the preliminary correction function F2 after generating a set of correction objective function F3 and the corresponding correction result function F2. Compare and gradually approach the ideal value.

In this second embodiment, after generating a plurality of sets of corresponding correction objective functions F3 and correction result functions F2, the corresponding correction objective functions F3 and correction result functions F2 are then provided to the deep learning system to execute all The training procedure of the deep learning system is described (step S304). Specifically, the adopted deep learning system may be a framework such as Caffe, Theano, TensorFlow, or Lasagne, Keras, or even DSSTNE, etc., and the present invention does not specifically limit which framework to adopt. Thereby, a prediction algorithm can be established by the deep learning system to predict the correction result function F2 corresponding to the input correction objective function F3 (step S306).

In specific applications, some of the combinations can be provided to a deep learning system to establish a prediction algorithm, and the remaining combinations can be used to verify the correctness of the prediction algorithm. For example, if there are 100 sets of correction objective function F3 and correction result function F2 corresponding to each other, 80 sets of correction objective function F3 and correction result function F2 corresponding to each other are provided to the deep learning system to execute the training program, and accordingly Build predictive algorithms. After the prediction algorithm is established, the remaining 20 groups of corresponding correction objective function F3 and correction result function F2 are used for verification. Among them, for the purpose of the present invention, the better verification method is through the prediction algorithm. The modified objective function F3 corresponding to the predicted modified result function F2 is compared with the previously known modified objective function F3 when the prediction result is obtained to confirm the results predicted by the prediction algorithm. Whether the result is consistent with the modified objective function F3 corresponding to the modified result function F2.

If the obtained prediction result is not satisfactory, a larger number of modified objective functions F3 can be generated by random numbers, and more sets of correspondences between the modified result functions F2 and the modified objective functions F3 can be established, and the deep learning system is executed again. Training procedures. Furthermore, you can also consider performing a random number operation based on a certain correction result function F2 (compared to the preliminary result function F2 used at the beginning) and the preliminary objective function F1, so as to generate more diverse correction result functions. The corresponding relationship between F2 and the modified objective function F3, so that the training program of the deep learning system can obtain more diverse learning samples.

Finally, when several verifications have been performed and the accuracy of the results predicted by the evaluation prediction algorithm can meet expectations, the preliminary objective function F1 obtained initially based on the image P1 corresponding to the target light shape can be set as a prediction The correction result function F2 in the algorithm, and the corresponding correction objective function F3 is inferred by the prediction algorithm (step S308). After obtaining the modified objective function F3 according to this, step S214 shown in FIG. 1 may be further performed to obtain a corresponding free-form solid model M, and the optical free-form surface structure required for production may be produced according to this free-form solid model M. . On the other hand, before actual production, step S216 shown in FIG. 1 may be further performed to actually simulate the correction result function F2 corresponding to the correction objective function F3 and confirm whether it is indeed consistent with the preliminary objective function F1. .

Based on the above description, with the flow shown in FIG. 9 and the component symbols shown in FIG. 2 to FIG. 8, the main flow structure for obtaining the finally required modified objective function F3 in the second embodiment of the present invention is briefly summarized as follows :

Step S300: the random number generates a large number of modified objective functions F3;

Step S302: According to steps 214 to S218 shown in FIG. 1, a correction result function F2 corresponding to each of the correction objective functions F3 is generated;

Step S304: Provide the corresponding correction target function F3 and the correction result function F2 to the deep learning system to execute the training program of the deep learning system.

Step S306: Establish a prediction algorithm through a deep learning system to predict the correspondence between the modified objective function F3 and the modified result function F2; and

Step S308: The preliminary objective function F1 is set to the correction result function F2, and the corresponding correction objective function F3 is inferred by the prediction algorithm.

[Parameter pre-adjustment]

In order to ensure that the steel result light trajectory L simulated by the beam tracing software can be closer to the light trajectory L in actual application, and then the most accurate result light shape is simulated, the beam tracing software needs to be targeted according to the material properties used, etc. Each setting parameter is fine-tuned. In the present invention, before the modified free-form surface solid model M is introduced into the beam tracking software, a step of pre-adjusting parameters is further included. Specifically, first, a plurality of different three-dimensional models are constructed first. The higher the diversity, the more it can reflect the change of the light trajectory L at different angles, and the better the effect in subsequent applications. After establishing a plurality of different three-dimensional models, according to the three-dimensional models, the materials and processes used to actually produce the optical free-form surface structure are used to prepare multiple optical structures for samples, and according to the multiple samples, The optical structure corrects the setting parameters of the beam tracking software.

Specifically, after preparing a plurality of sample optical structures with the materials to be used and the manufacturing process, the prepared samples are actually illuminated with the optical structures to obtain the actual light trajectory L and the actual resulting light shape. In other words, firstly, the optical structures for a plurality of samples are passed through to generate a plurality of corresponding illumination patterns, respectively. Then, after obtaining the actual light trajectory L and the actual formed light shape, the parameters in the beam tracing software can be adjusted by comparing the results simulated by the beam tracing software with the actual results. That is, a plurality of three-dimensional models are respectively introduced into the beam tracing software to generate a plurality of simulated sample light shapes corresponding to the plurality of three-dimensional models, respectively. After analyzing the previously obtained multiple light patterns, multiple actual light shapes corresponding to the multiple light patterns can be obtained. By comparing multiple actual light shapes with corresponding multiple simulated sample light shapes, the beam tracking software can be modified accordingly. Setting parameters.

On the other hand, in order to make the simulated free-form three-dimensional model M and the optical free-form surface structure actually produced have better fidelity, the present invention also introduces the modified free-form three-dimensional model M into the beam tracking software before The optical structure for multiple samples can be used to correct the physical property data and molding condition parameters required when performing mold flow analysis. In specific applications, CAV technology can be used to perform reverse three-dimensional surface scanning technology to confirm the actual shape of the optical structure for multiple samples, and compare the multiple actual shapes with corresponding corresponding three-dimensional models, which can be corrected accordingly. Physical property data and molding condition parameters required for mold flow analysis.

[Advantageous Effects of the Embodiment]

One of the beneficial effects of the present invention is that the method for making a one-time optical free-form surface structure provided by the present invention can obtain a value between one of the two by “comparing a preset objective function with a correction result function calculated by simulation”. "Difference value" and "repeatedly adjust the modified objective function when the difference value has not fallen below a preset threshold" to obtain the best modified free-form solid model and make a primary optical that can produce the desired target light shape Free-form surface structure.

The contents disclosed above are only the preferred and feasible embodiments of the present invention, and therefore do not limit the scope of patent application of the present invention. Therefore, any equivalent technical changes made by using the description and drawings of the present invention are included in the application of the present invention Within the scope of the patent.

Claims (10)

A manufacturing method of a primary optical free-form surface structure includes: (A) obtaining an image corresponding to a target light shape; (B) analyzing the image to obtain a preliminary target corresponding to the target light shape Function; (C) performing a free-form surface calculation of a point light source according to the preliminary objective function to obtain a preliminary free-form solid model; (D) importing the preliminary free-form solid model into a beam tracing software to simulate a Preliminary result light shape; (E) analyzing the preliminary result light shape to obtain a preliminary result function corresponding to the preliminary result light shape; (F) comparing the preliminary result function and the preliminary objective function; (G) ) Generating a modified objective function according to a comparison result between the preliminary result function and the preliminary objective function; (H) performing a free-form surface operation of a volume light source according to the modified objective function to obtain a modified free-surface solid model (I) importing the modified free-form surface solid model into the beam tracing software to simulate a correction result light shape; (J) analyzing the correction result light shape to obtain a corresponding correction result A correction result function of the shape; (K) comparing the correction result function and the preliminary objective function to obtain a difference value between the two; (L) determining whether the difference value is lower than a preset threshold value (M) when the difference value is not lower than the preset threshold value, according to a comparison result between the correction result function and the preliminary objective function, to generate another modified objective function, and repeat step (H) To (L); and (N) when the difference value is lower than the preset threshold value, according to the modified free-form solid model, to make the primary optical free-form structure. The manufacturing method according to claim 1, wherein the process of generating the modified objective function further includes the following steps: on the premise that the area under the curve of the preliminary objective function is maintained constant, adjusting all the factors according to the preliminary result function. The preliminary objective function is described. The manufacturing method according to claim 2, wherein when the value of the preliminary result function is higher than the coordinate position of the value of the preliminary objective function, the value of the preliminary objective function is reduced. The manufacturing method according to claim 2, wherein when the value of the preliminary result function is lower than the coordinate position of the value of the preliminary objective function, the value of the preliminary objective function is raised to generate the correction target. function. The manufacturing method according to claim 1, wherein the process of generating the modified objective function further includes the following steps: generating a large number of the modified objective function according to a difference function, and according to steps (H) to ( J) generating the correction result function corresponding to each of the correction objective functions; providing the correction objective function and the correction result function corresponding to each other to a deep learning system to perform training of the deep learning system A program; establishing a prediction algorithm through the deep learning system, the prediction algorithm being used to predict the correction result function corresponding to the input correction objective function; and setting the preliminary objective function to the The result function is modified, and the modified objective function is inferred by the prediction algorithm. The method according to claim 1, wherein before introducing the modified free-form surface solid model into the beam tracing software, the method further comprises the following steps: constructing a plurality of solid models; and separately preparing a plurality of solid models based on the plurality of solid models. An optical structure for each sample; and a plurality of optical structures for the sample to modify the setting parameters of the beam tracking software. The manufacturing method according to claim 6, wherein in the process of correcting the setting parameters of the beam tracing software by using a plurality of optical structures for the sample, the method further includes the following steps: passing a plurality of optical structures for the sample Generating a plurality of corresponding lighting patterns respectively; importing a plurality of the three-dimensional models into the beam tracing software to generate a plurality of simulated sample light shapes respectively corresponding to a plurality of the three-dimensional models; and analyzing a plurality of the light patterns A plurality of actual light shapes of the illumination pattern, and comparing the plurality of actual light shapes with the corresponding plurality of simulated sample light shapes, respectively, to modify the setting parameters of the light beam tracking software. The method according to claim 1, wherein before introducing the modified free-form surface solid model into the beam tracing software, the method further comprises the following steps: constructing a plurality of solid models; and separately preparing a plurality of solid models based on the plurality of solid models. An optical structure for each sample; and correcting physical property data and molding condition parameters required for performing a mold flow analysis by using a plurality of the optical structures for the sample. The manufacturing method according to claim 8, wherein the process of correcting physical property data and molding condition parameters required for performing the mold flow analysis by using a plurality of the optical structures of the sample further includes the following steps: The three-dimensional surface scanning technology confirms an actual shape of a plurality of the optical structures for the sample, and compares the plurality of actual shapes with a corresponding plurality of the three-dimensional models, respectively, so as to modify the position when performing the mold flow analysis. Required physical property data and molding condition parameters. The manufacturing method according to claim 8, further comprising: performing the mold flow analysis on the modified free-form surface solid model to establish a mold required for producing the primary optical free-form surface structure.