Computer program, lens layout generation apparatus, lens layout generation method, and learning model generation method

Abstract

To provide a computer program, a lens layout generation apparatus, a lens layout generation method and a learning model generation method that are capable of generating layout data of a lens group that is obtained by combining a plurality of lenses listed in a lens catalog.SOLUTION: A computer program causes a computer to execute a process in which a field of view, a numerical aperture and an effective focal length of a lens group formed by combining a plurality of lenses are acquired, and, when the field of view, the numerical aperture and the effective focal length are input, the acquired field of view, numerical aperture and effective focal length are input into a learning model that generates layout data of a plurality of lenses recorded in a lens catalog database, thereby generating layout data of the lens group.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a computer program, a lens layout generation device, a lens layout generation method, and a learning model generation method. [Background technology]

[0002] Designing an optical layout is a complex task, and is usually performed by an experienced designer with specialized knowledge of optical layout using ray tracing software, etc. Depending on the specifications of the optical system and the target cost, the designer lays out the lenses according to a predefined lens configuration while minimizing image distortion within the field of view and aperture range of the optical system.

[0003] Many methods have been proposed to automate this design process. Patent Document 1 discloses a method for designing an ideal shape of a lens surface from a set of object points and image points using a ray tracing method. [Prior art documents] [Non-patent literature]

[0004] [Non-Patent Document 1] “Automated design of freeform imaging system”, Tong Yang, Guo-Fan Jin and Jun Zhu, Light: Science & Applications (2017) Summary of the Invention [Problem to be solved by the invention]

[0005] However, the technique of Patent Document 1 cannot generate a lens layout, such as the number and type of lenses that make up the optical system, or the relative positions of the lenses. Furthermore, since the lenses that make up the optical system are custom products with newly designed surface shapes, the cost of the optical system tends to be high and the delivery time for the lenses tends to be longer than for standard products.

[0006] The present invention has been made in consideration of the above circumstances, and aims to provide a computer program, a lens layout generation device, a lens layout generation method, and a learning model generation method that are capable of generating layout data for a lens group that combines multiple lenses listed in a lens catalog. [Means for solving the problem]

[0007] The present application includes multiple means for solving the above-described problems. As one example, a computer program causes a computer to execute a process of acquiring the field of view, numerical aperture, and effective focal length of a lens group formed by combining multiple lenses, and, when the field of view, numerical aperture, and effective focal length are input, inputting the acquired field of view, numerical aperture, and effective focal length into a learning model that generates layout data of multiple lenses recorded in a lens catalog database, thereby generating layout data of the lens group. [Effects of the Invention]

[0008] According to the present invention, layout data for a lens group that combines a plurality of lenses listed in a lens catalog can be generated. [Brief explanation of the drawings]

[0009] [Figure 1] FIG. 1 is a diagram illustrating an example of the configuration of a lens layout generating device according to a first embodiment. [Figure 2] 1 is a diagram showing an example of the field of view (FOV), numerical aperture (NA), and light loss of a lens group. [Figure 3] FIG. 10 is a diagram showing an example of layout data of a lens group. [Figure 4] FIG. 2 is a diagram illustrating an example of the configuration of a lens catalog database. [Figure 5] FIG. 10 is a diagram showing an example of a method for generating layout data using a second learning model. [Figure 6] FIG. 2 is a diagram illustrating an example of a method for generating learning data for a second learning model and a method for generating (learning) the second learning model in the first embodiment. [Figure 7] FIG. 10 is a diagram showing an example of a procedure for generating layout data by the lens layout generating device. [Figure 8] FIG. 2 is a diagram illustrating an example of a method for generating a first learning model and a second learning model. [Figure 9] FIG. 10 is a diagram illustrating an example of the configuration of a lens layout generating device according to a second embodiment. [Figure 10] FIG. 10 is a diagram illustrating an example of a method for generating learning data for a second learning model and a method for generating (learning) the second learning model in the second embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0010] (First embodiment) The present invention will be described below with reference to the drawings illustrating embodiments thereof. FIG. 1 is a diagram showing an example of the configuration of a lens layout generation device 50 according to a first embodiment. The lens layout generation device 50 includes a control unit 51 that controls the entire device, an input unit 52, a memory 53, a display unit 54, an operation unit 55, an interface unit 56, a determination unit 57, a ray tracing unit 58, a learning processing unit 59, and a storage unit 60. A lens catalog database 100 is connected to the lens layout generation device 50. The lens catalog database 100 may be configured as a data server or the like. The lens catalog database 100 can be configured as a computer (PC) or the like.

[0011] The control unit 51 may be configured by incorporating a required number of central processing units (CPUs), micro-processing units (MPUs), graphics processing units (GPUs), etc. The control unit 51 may also be configured by combining digital signal processors (DSPs), field-programmable gate arrays (FPGAs), etc.

[0012] The input unit 52 receives input data for generating layout data of the lens group.

[0013] The display unit 54 can be configured with a liquid crystal panel, an organic EL (Electro Luminescence) display, or the like.

[0014] The operation unit 55 is configured with a touch panel or the like, and can accept operations on the information displayed on the display unit 54. The operation unit 55 may also be configured with a keyboard, a mouse, a touch pad, or the like.

[0015] The interface unit 56 has an interface function for accessing the lens catalog database 100. The interface unit 56 writes data to the lens catalog database 100 and reads data from the lens catalog database 100.

[0016] The storage unit 60 can be configured with, for example, a hard disk or semiconductor memory, and stores a computer program (program product) 61, a first learning model 62, a second learning model 63, and required information.

[0017] The memory 53 can be configured with semiconductor memory such as SRAM (Static Random Access Memory), DRAM (Dynamic Random Access Memory), flash memory, etc. A computer program 61 can be loaded into the memory 53, and the control unit 51 can execute the computer program 61.

[0018] The computer program 61 can be stored in the storage unit 60 by reading a storage medium M (for example, an optically readable disk storage medium such as a CD-ROM) on which the computer program 61 is recorded using a storage medium reading unit (not shown). The computer program 61 can also be downloaded from an external device via the input unit 52 and stored in the storage unit 60.

[0019] The learning processing unit 59 performs the processes of learning, relearning, and tuning (collectively referred to as "generation") the first learning model 62 and the second learning model 63.

[0020] The first learning model 62 is trained (generated) so that when layout data of a lens group is input, it outputs estimated values ​​of the field of view (FOV) and numerical aperture (NA) of the lens group. Furthermore, when layout data of a lens group is input, the first learning model 62 is trained so that it outputs estimated values ​​of the optical loss of the lens group. The first learning model 62 may be configured, for example, with a deep neural network (DNN) or a recurrent neural network (RNN). Initial training of the first learning model 62 may be performed using design data from an existing design database (not shown). In other words, the first learning model 62 is a model for which initial training has been performed using an existing design database (unrelated to the lens catalog database 100).

[0021] The second learning model 63 is trained (generated) to generate layout data of a plurality of lenses recorded in the lens catalog database 100 when the field of view, numerical aperture, and effective focal length (EFL) of the lens group are input. Furthermore, the second learning model 63 is trained to generate layout data of a plurality of lenses recorded in the lens catalog database 100 when at least one of the price and delivery date of the lens group is further input. The second learning model 63 can be configured, for example, by a DNN (Deep Neural Network) or an RNN (Recurrent Neural Network).

[0022] The determination unit 57 determines whether the estimated values ​​of the field of view, numerical aperture, and optical loss output by the first learning model 62 are within an allowable range. Whether they are within an allowable range means whether the layout data of the lens group input to the first learning model 62 is feasible or whether a lens group according to the layout data can actually be configured.

[0023] If the determination unit 57 determines that the estimated value is within the allowable range, the ray tracing unit 58 applies a ray tracing method to the arrangement of each lens in the lens group according to the lens group layout data input to the first learning model 62, thereby calculating the field of view and numerical aperture of the lens group. Note that the field of view and numerical aperture calculated by the ray tracing unit 58 are significantly more realistic than the estimated values ​​of field of view and numerical aperture output by the first learning model 62. The ray tracing method can trace rays using a set of object points and image points corresponding to the object points, and calculate the field of view, numerical aperture, and effective focal length based on the traced rays. Note that ray tracing, which calculates the path of light traveling through an optical system based on geometry, is not limited to the so-called (n,V) method, which uses the radius of curvature of the lens, the aperture radius, the refractive index, the z coordinate of the center position of the lens surface, etc., and other methods (polynomial coefficients) such as the Sellmeier method (the Sellmeier dispersion formula using the dispersion formula showing the relationship between refractive index and wavelength) may also be used.

[0024] Figure 2 shows an example of the field of view (FOV), numerical aperture (NA), and light loss of a lens group. As shown in Figure 2A, the field of view (FOV) is the maximum area of ​​a scene that can be captured by a sensor (imaging surface) through the lens group. The field of view of a lens group can be calculated using the angle of view (AFOV), the width of an object (target object) visible through the lens group (HFOV), and the distance between the lens group and the object, but in most cases, the field of view can be expressed using the AFOV or HFOV.

[0025] As shown in Figure 2B, the numerical aperture (NA) can be calculated as NA = n·sinθ, where n is the refractive index of the lens and θ is the maximum angle at which light can enter the lens group from the object. The larger the numerical aperture, the wider the area of ​​light that can be collected.

[0026] As shown in Figure 2C, light loss is the amount of light incident on the lens group that cannot be collected by the sensor. The greater the difference between the distance between the lens group and the sensor and the focal length of the lens group, the greater the light loss. The light loss function can be expressed as follows: Light Loss Function = {W1 × Spot Size + W2 × Ray Loss Function}. Spot Size is the area (point spread function) created by a ray of light tracing from a point on the input light source through the optical system. Ray loss corresponds to the number of rays that are not transmitted to the screen, for example, due to lens diameter or insufficient total reflection. Ray loss occurs not only between the final lens and the screen, but also between lenses. The training rate of the first learning model 62 can be improved by adjusting the coefficients W1 and W2.

[0027] FIG. 3 shows an example of layout data for a lens group. For convenience, let us assume that the lens group is composed of three lenses: L1, L2, and L3. Lenses L1 and L3 are biconvex lenses, and lens L2 is a biconcave lens. Let us assume that light travels through lenses L1, L2, and L3 in this order. Let us assume that the radius of curvature of the entrance surface of lens L1 is R11, and the radius of curvature of the exit surface is R12. Let us assume that the radius of curvature of the entrance surface of lens L2 is R21, and the radius of curvature of the exit surface is R22. Let us assume that the radius of curvature of the entrance surface of lens L3 is R31, and the radius of curvature of the exit surface is R32. Let us assume that the separation dimension between lenses L1 and L2 is D1, and the separation dimension between lenses L2 and L3 is D2. Let us assume that the refractive indices of lenses L1, L2, and L3 are n1, n2, and n3, and the Abbe numbers of lenses L1, L2, and L3 are V1, V2, and V3. The Abbe number is an index used to evaluate the chromatic dispersion of transparent materials (change in refractive index with wavelength). The refractive index and Abbe number can completely characterize the lens material.

[0028] In this case, the layout data of the lens group can be expressed as {L1, L2, L3, R11, R12, D1, R21, R22, D2, R31, R32, n1, n2, n3, V1, V2, V3}. That is, the lens layout data can be expressed in a format such as {arrangement order of each lens constituting the lens group and lens ID, radius of curvature of the first lens, distance between the first and second lenses, radius of curvature of the second lens, distance between the second and third lenses, ..., refractive index of each lens, Abbe number of each lens}. Note that the format of the layout data is not limited to the example of FIG. 3. Also, the configuration of the lens group is merely an example and is not limited to the example of FIG. 3.

[0029] FIG. 4 is a diagram showing an example of the configuration of the lens catalog database 100. The lens catalog database 100 has catalog data such as that shown in FIG. 4A and training data such as that shown in FIG. 4B. The catalog data includes information on standard and commercially available lenses as listed in lens manufacturer catalogs. For example, the catalog data includes information such as the lens ID (product number, etc.), type (e.g., biconvex lens, plano-convex lens, convex menisca lens, biconcave lens, plano-concave lens, concave menisca lens, etc.), lens size (radius or diameter) and thickness, radius of curvature of the lens surface, focal length, refractive index, price, and delivery time. Note that the catalog data may (1) not include information on price and delivery time, (2) include information on price, (3) include information on delivery time, or (4) include information on price and delivery time.

[0030] As mentioned above, the lens catalog database 100 may include the type of lens and the radius of curvature of the lens surface, as well as the price and / or delivery time of the lens.

[0031] The training data includes data in which the field of view (FOV), numerical aperture (NA), effective focal length (EFL), price, delivery time, and layout data of the lens group are associated with each lens group (combination of lenses). Different training data is associated with different numbers of lenses in the lens group, different lens types, and different arrangements between lenses. The effective focal length (EFL) is the effective focal length of the lens group (optical system).

[0032] 5 is a diagram showing an example of a method for generating layout data using the second learning model 63. The control unit 51 acquires the field of view, numerical aperture, and effective focal length of a lens group formed by combining multiple lenses, and when the field of view, numerical aperture, and effective focal length are input, the control unit 51 inputs the acquired field of view, numerical aperture, and effective focal length into the second learning model 63, which generates layout data for multiple lenses recorded in the lens catalog database 100, to generate layout data for the lens group.

[0033] In addition, the control unit 51 acquires at least one of the price and delivery date of the lens group, and if at least one of the price and delivery date is further input, it may input at least one of the acquired price and delivery date into a second learning model 63 that generates layout data of multiple lenses recorded in the lens catalog database 100, thereby generating layout data of the lens group.

[0034] In the example of FIG. 5, when {FOV1, NA1, EFL1, price1, delivery date1} is input as input data to the second learning model 63, the second learning model 63 generates and outputs layout data {RA1}. Also, when {FOV2, NA2, EFL2, price2, delivery date2} is input as input data to the second learning model 63, the second learning model 63 generates and outputs layout data {RA2}. Similarly, when {FOVn, NAn, EFLn, pricen, delivery daten} is input as input data to the second learning model 63, the second learning model 63 generates and outputs layout data {RAn}. In either case, the second learning model 63 can generate layout data that minimizes optical loss. In the example of FIG. 5, the price and delivery date may be omitted, or at least one of them may be input, or both the price and delivery date may be input.

[0035] As described above, according to this embodiment, layout data for a lens group combining multiple lenses listed in a lens catalog can be generated simply by inputting at least the field of view, numerical aperture, and effective focal length. Furthermore, layout data for a lens group combining multiple lenses listed in a lens catalog can be generated simply by further inputting at least one of the price and delivery time. This allows for the acquisition of layout data for a lens group combining commercially available (not custom-made) lenses with clear prices and delivery times, making it easy to design an optical layout.

[0036] Next, a learning method for the second learning model 63 using the catalog data recorded in the lens catalog database 100 will be described.

[0037] 6 is a diagram showing an example of a method for generating learning data for the second learning model 63 in the first embodiment and a method for generating (learning) the second learning model 63. The entire process shown in FIG. 6 can be controlled by the control unit 51. First, the method for generating learning data will be described.

[0038] The learning processing unit 59 selects a plurality of lenses recorded in the catalog data from the lens catalog database 100. Each of the plurality of lenses includes information such as a lens ID, a lens type, and a radius of curvature of the lens surface (see FIG. 4A).

[0039] The learning processing unit 59 determines layout data (see FIG. 3) of a lens group combining the selected lenses, and outputs the determined layout data to the first learning model 62 and the determination unit 57. The layout data can be determined by appropriately setting the order in which the lenses are arranged, the dimensions between the lenses, and the like.

[0040] When layout data of a lens group is input, the first learning model 62 generates estimated values ​​of the field of view (FOV) and numerical aperture (NA) of the lens group and outputs the generated estimated values ​​to the determination unit 57. In addition, when layout data of a lens group is input, the first learning model 62 may generate an estimated value of the optical loss of the lens group and output the generated estimated value to the determination unit 57.

[0041] Based on the estimated values ​​output by the first learning model 62, the determination unit 57 determines whether the layout of the lens group is feasible, i.e., whether there are actually any practical problems with the lens group. Specifically, thresholds that define the allowable ranges of the field of view, numerical aperture, and light loss are set, and the determination unit 57 determines whether the estimated values ​​are within the allowable ranges. If the estimated values ​​are not within the allowable ranges, the layout of the lens group is discarded as a realization load. If the estimated values ​​are within the allowable ranges, the determination unit 57 outputs layout data of the lens group to the ray tracing unit 58 and the learning processing unit 59.

[0042] The ray tracing unit 58 lays out the lenses based on the layout data of the lens group, calculates the progress of rays between object points (points on the object) and image points (points on the sensor) using ray tracing, and calculates the field of view, numerical aperture, and effective focal length (EFL) of the lens group. The ray tracing unit 58 outputs the calculated field of view, numerical aperture, and effective focal length to the learning processing unit 59.

[0043] The learning processing unit 59 generates learning data by associating the input field of view, numerical aperture, and effective focal length with layout data, and stores the generated learning data in the lens catalog database 100. Note that the learning data may also be created by associating the input field of view, numerical aperture, and effective focal length with at least one of the price and delivery date of the selected lens group and layout data. The price of the lens group may be the sum of the prices of the lenses that make up the lens group. The delivery date of the lens group may be the longest delivery date among the delivery dates of the lenses that make up the lens group.

[0044] The control unit 51 can repeatedly execute the above-described process until a required amount of learning data is collected.

[0045] Through the above-described processing, it is possible to collect, as learning data, layout data of lens groups that combine standard lenses and commercially available lenses (not custom lenses) stored in the lens catalog database 100. Furthermore, the learning data is associated with the price and delivery time of the lens groups, so it is possible to know the price and acquisition time of the lens group for which the layout is designed, and also to acquire lenses at a lower price and with a shorter delivery time than custom lenses.

[0046] Next, the learning method of the second learning model 63 will be described.

[0047] The learning processing unit 59 can acquire learning data from the lens catalog database 100 and perform learning model generation processes such as learning, relearning, and tuning of the second learning model 63 using the acquired learning data. Specifically, the learning processing unit 59 can generate the second learning model based on the learning data so that, when a field of view, a numerical aperture, and an effective focal length are input, layout data of a lens group (plurality of lenses) associated with the input field of view, numerical aperture, and effective focal length is generated. Furthermore, the learning processing unit 59 may generate the second learning model based on the learning data so that, when a field of view, numerical aperture, and effective focal length, and at least one of a price and a delivery date are input, layout data of a lens group (plurality of lenses) associated with the input field of view, numerical aperture, and effective focal length, and at least one of a price and a delivery date is generated.

[0048] The first learning model 62 can also be trained using learning data. That is, when layout data of a lens group (plurality of lenses) is input based on the learning data, the learning processing unit 59 can generate the first learning model 62 so as to output a field of view and a numerical aperture associated with the input layout data. Note that by performing initial training of the first learning model 62 using design data from an existing design database, the learning processing unit 59 can update the first learning model 62 based on the learning data.

[0049] Furthermore, when the learning processing unit 59 receives layout data of a lens group, it can generate the first learning model 62 so as to output an estimated value of the optical loss of the lens group.

[0050] 7 is a diagram showing an example of the procedure for generating layout data by the lens layout generating device 50. For convenience, the following description will be given with the control unit 51 as the main actor in the processing. The control unit 51 acquires the field of view (FOV), numerical aperture (NA), and effective focal length (EFL) of the lens group, as well as at least one of price and delivery date (S11), and inputs the acquired field of view, numerical aperture, and effective focal length, as well as at least one of price and delivery date into the second learning model 63 (S12). The control unit 51 generates layout data of the lens group based on the output of the second learning model 63 (S13), and ends the processing.

[0051] The layout data that is generated is layout data of a lens group that combines standard lenses and commercially available lenses (not custom lenses) stored in the lens catalog database 100, so compared to generating layout data using custom lenses, a layout-designed lens group can be obtained at a lower price and in a shorter delivery time.

[0052] 8 is a diagram showing an example of a learning method for the first learning model 62 and the second learning model 63. The learning method includes re-learning, tuning, generation method, etc. The control unit 51 determines layout data for multiple lenses selected from the lens catalog database 100 (S21), and inputs the determined layout data into the first learning model 62 (S22).

[0053] The control unit 51 determines whether the field of view, numerical aperture, and optical loss (estimated values) of the lens group formed by combining the multiple lenses output by the first learning model 62 are within the acceptable range (S23), and if the estimated values ​​are within the acceptable range (YES in S23), it applies a ray tracing method to the determined layout data to calculate the field of view, numerical aperture, and effective focal length of the lens group (S24).

[0054] The control unit 51 associates the determined layout data with the calculated field of view, numerical aperture, and effective focal length and stores them as learning data in the lens catalog database 100 (S25), and determines whether a required amount of learning data has been collected (S26). If a required amount of learning data has been collected (YES in S26), the control unit 51 uses the learning data to train the first learning model 62 and the second learning model 63 (S27), and ends the process.

[0055] If the estimated value is not within the allowable range (NO in S23), or if the required amount of learning data has not been collected (NO in S26), the control unit 51 continues the processing from step S21 onwards.

[0056] In the present embodiment, the lens layout generation device 50 has a configuration including the determination unit 57, the ray tracing unit 58, and the learning processing unit 59, but is not limited to this. For example, the determination unit 57, the ray tracing unit 58, and the learning processing unit 59 may be provided in a separate external server or cloud.

[0057] (Second embodiment) 9 is a diagram showing an example of the configuration of a lens layout generation device 50 according to the second embodiment. The difference from the first embodiment shown in FIG. 1 is that a random lens generator 64 and a custom lens database 200 are provided.

[0058] The custom lens database 200 includes information on lenses (custom products) that have been customized based on user requests, etc. Similar to the lens catalog database 100, the custom lens database 200 includes information such as the lens ID, lens type, radius of curvature of the lens surface, focal length, refractive index, and Abbe number.

[0059] The random lens generator 64 can randomly select and freely combine lenses from among the lenses recorded in the lens catalog database 100 and the lenses recorded in the custom lens database 200. The random lens generator 64 can identify whether the selected lens is a lens recorded in the lens catalog database 100 or a lens recorded in the custom lens database 200.

[0060] 10 is a diagram showing an example of a method for generating learning data for the second learning model 63 of the second embodiment and a method for generating (learning) the second learning model 63. The difference from the first embodiment shown in FIG. 6 is that the output of the random lens generator 64 is input to the learning processing unit 59.

[0061] The random lens generator 64 accesses the lens catalog database 100 and the custom lens database 200 to determine which lenses to extract from the lens catalog database 100 and which lenses to extract from the custom lens database 200. In this case, extraction source information is specified to indicate which of the extracted lenses were extracted from the lens catalog database 100 and which lenses were extracted from the custom lens database 200. The random lens generator 64 also obtains information about the extracted lenses from the respective databases. The random lens generator 64 outputs the extraction source information about the extracted lenses (e.g., whether they are catalog products or custom products) and information about the lenses (e.g., field of view, numerical aperture, and effective focal length) to the learning processing unit 59 as learning data.

[0062] The learning processing unit 59 can acquire learning data from the random lens generator 64 and perform learning model generation processes such as learning, relearning, and tuning of the second learning model 63 using the acquired learning data. Specifically, as in the first embodiment, the learning processing unit 59 can generate the second learning model based on the learning data so that, when a field of view, a numerical aperture, and an effective focal length are input, layout data of a lens group (plurality of lenses) associated with the input field of view, numerical aperture, and effective focal length is generated. Furthermore, when a field of view, a numerical aperture, an effective focal length, and at least one of a price and a delivery date are input, the learning processing unit 59 can generate the second learning model so that layout data of a lens group (plurality of lenses) associated with the input field of view, numerical aperture, effective focal length, and at least one of a price and a delivery date is generated based on the learning data. The number of custom lenses can also be included in the learning data.

[0063] The above-described configuration allows not only catalog lenses but also custom lenses to be freely combined, thereby increasing the degree of freedom in constructing a good optical system.

[0064] (Supplementary Note 1) The computer program causes a computer to execute a process of acquiring the field of view, numerical aperture, and effective focal length of a lens group combining multiple lenses, and inputting the acquired field of view, numerical aperture, and effective focal length into a learning model that generates layout data of multiple lenses recorded in a lens catalog database when the field of view, numerical aperture, and effective focal length are input, thereby generating layout data of the lens group.

[0065] (Supplementary Note 2) The computer program in Supplementary Note 1 acquires at least one of the price and delivery date of the lens group, and when at least one of the price and delivery date is further input, causes the computer to execute a process of inputting the acquired at least one of the price and delivery date into the learning model that generates layout data of a plurality of lenses recorded in a lens catalog database, thereby generating layout data of the lens group.

[0066] (Supplementary Note 3) In the computer program according to Supplementary Note 1 or Supplementary Note 2, the lens catalog database includes at least one of the lens type, the radius of curvature of the lens surface, and the price and delivery time of the lens.

[0067] (Appendix 4) The lens layout generation device includes a control unit, which acquires the field of view, numerical aperture, and effective focal length of a lens group combining multiple lenses, and when the field of view, numerical aperture, and effective focal length are input, inputs the acquired field of view, numerical aperture, and effective focal length into a learning model that generates layout data of multiple lenses recorded in a lens catalog database, thereby generating layout data of the lens group.

[0068] (Appendix 5) The lens layout generation method acquires the field of view, numerical aperture, and effective focal length of a lens group combining multiple lenses, and when the field of view, numerical aperture, and effective focal length are input, inputs the acquired field of view, numerical aperture, and effective focal length into a learning model that generates layout data of multiple lenses recorded in a lens catalog database, thereby generating layout data of the lens group.

[0069] (Appendix 6) A learning model generation method determines layout data of multiple lenses selected from a lens catalog database including the lens type and the radius of curvature of the lens surface, inputs the determined layout data into a first learning model that, when inputting layout data of a lens group, outputs estimated values ​​of the field of view and numerical aperture of the lens group, obtains estimated values ​​of the field of view and numerical aperture of the lens group combining the multiple lenses, determines whether the obtained estimated values ​​are within an acceptable range, and if it is determined that the estimated values ​​are within an acceptable range, calculates the field of view, numerical aperture, and effective focal length of the lens group combining the multiple lenses using a ray tracing method, stores learning data in the lens catalog database that corresponds the calculated field of view, numerical aperture, and effective focal length to the layout data of the multiple lenses, and generates a second learning model based on the learning data so that, when inputting a field of view, numerical aperture, and effective focal length, it generates layout data of the multiple lenses that corresponds to the input field of view, numerical aperture, and effective focal length.

[0070] (Appendix 7) The learning model generation method in Appendix 6 generates the first learning model based on the learning data so as to output a field of view and a numerical aperture corresponding to the input layout data when layout data of the plurality of lenses is input.

[0071] (Supplementary Note 8) The learning model generation method according to Supplementary Note 6 or Supplementary Note 7 generates the first learning model so as to output an estimated value of optical loss of the lens group when layout data of the lens group is input.

[0072] The matters described in each embodiment can be combined with each other. In addition, the independent claims and dependent claims described in the claims can be combined with each other in any combination, regardless of the reference format. Furthermore, although the claims use a format in which a claim references two or more other claims (multiple claim format), this is not limited to this format. A multiple claim (multi-multi claim) that references at least one other multiple claim may also be used. [Explanation of symbols]

[0073] 50 Lens layout generator 51 Control section 52 Input section 53 Memory 54 Display section 55 Operation section 56 Interface section 57 Judgment section 58 Ray Tracing Department 59 Learning processing unit 60 Storage section 61 Computer Programs 62 First Learning Model 63 Second Learning Model 64 Random Lens Generator 100 Lens Catalog Database 200 custom lens database

Claims

1. Obtain the field of view, numerical aperture, and effective focal length of a lens group that combines multiple lenses, inputting the acquired field of view, numerical aperture, and effective focal length into a learning model that generates layout data of a plurality of lenses recorded in a lens catalog database when the field of view, numerical aperture, and effective focal length are input, and generating layout data of the lens group; A computer program that causes a computer to perform a process.

2. Obtaining at least one of the price and delivery date of the lens group; When at least one of price and delivery date is further input, the acquired at least one of price and delivery date is input into the learning model that generates layout data of a plurality of lenses recorded in a lens catalog database, thereby generating layout data of the lens group.

2. The computer program according to claim 1, which causes a computer to execute a process.

3. The lens catalog database includes at least one of the lens type, the radius of curvature of the lens surface, and the price and delivery time of the lens.

3. A computer program according to claim 2.

4. A control unit is provided, The control unit Obtain the field of view, numerical aperture, and effective focal length of a lens group that combines multiple lenses, inputting the acquired field of view, numerical aperture, and effective focal length into a learning model that generates layout data of a plurality of lenses recorded in a lens catalog database when the field of view, numerical aperture, and effective focal length are input, and generating layout data of the lens group; Lens layout generator.

5. Obtain the field of view, numerical aperture, and effective focal length of a lens group that combines multiple lenses, inputting the acquired field of view, numerical aperture, and effective focal length into a learning model that generates layout data of a plurality of lenses recorded in a lens catalog database when the field of view, numerical aperture, and effective focal length are input, and generating layout data of the lens group; Lens layout generation method.

6. determining layout data of a plurality of lenses selected from a lens catalog database including lens types and lens surface curvature radii; When layout data of a lens group is input, the determined layout data is input to a first learning model that outputs estimated values ​​of a field of view and a numerical aperture of the lens group, thereby obtaining estimated values ​​of a field of view and a numerical aperture of the lens group combining the plurality of lenses; determining whether the obtained estimate is within an acceptable range; If it is determined that the values ​​are within the allowable range, a ray tracing method is used to calculate the field of view, the numerical aperture, and the effective focal length of the lens group formed by combining the plurality of lenses; storing learning data in the lens catalog database, the learning data associating the calculated field of view, numerical aperture, and effective focal length with layout data of the plurality of lenses; generating a second learning model based on the learning data so that, when a field of view, a numerical aperture, and an effective focal length are input, layout data of the plurality of lenses associated with the input field of view, numerical aperture, and effective focal length are generated; Learning model generation method.

7. generating the first learning model based on the learning data so as to output a field of view and a numerical aperture associated with the input layout data when layout data of the plurality of lenses is input; The learning model generation method according to claim 6 .

8. generating the first learning model so as to output an estimated value of optical loss of the lens group when layout data of the lens group is input; The learning model generation method according to claim 6 or 7.

Images