Optical element production data processing method and system

Abstract

The present application provides a method and system for processing optical element production data, which relates to optical element detection technology. In this method, the electronic device inputs the thickness data of the optical element to be processed and the process parameter data of the optical element processing area uploaded by the sensor device into a preset optical element quality prediction model for prediction processing, and obtains the predicted thickness data of the optical element to be processed when the processing in the optical element processing area is completed, so as to determine whether the optical element to be processed will have quality problems when it is processed in the optical element processing area. By applying this method, the quality of optical elements can be effectively improved.

Description

Technical Field

[0001] The present application relates to optical component detection technology, and in particular to an optical component production data processing method and system. Background Art

[0002] During the manufacturing process of optical components, quality control is a critical step in ensuring component performance.

[0003] In the existing technical solutions, optical component quality treatment technologies mainly include cleaning, grinding, polishing and coating technologies. Specifically, cleaning technologies include immersion cleaning, mechanical swing cleaning, solvent stirring cleaning, shower cleaning, steam cleaning and ultrasonic cleaning, etc., which aim to remove pollutants and impurities on the surface of optical components. Grinding and polishing technologies can remove scratches, cracks and other defects on the surface of optical components, and improve the surface finish and flatness. Coating technologies include physical vapor deposition, chemical vapor deposition, electrochemical deposition and other methods, which aim to form a uniform and dense thin film layer on the surface of optical components, thereby improving the optical performance and service life of optical components.

[0004] However, existing optical component quality processing technologies still have certain limitations in improving the quality of optical components. Summary of the invention

[0005] The present application provides an optical element production data processing method and system to solve the limitations faced by the existing technical solutions in improving the quality of optical elements.

[0006] In a first aspect, the present application provides a method for processing optical element production data, comprising:

[0007] Based on the data uploaded by the sensor device, the thickness data of the optical element to be processed and the process parameter data of the optical element processing area are obtained, wherein the process parameter data include the temperature data, humidity data, pressure data of the optical element processing area, and the pressure data and optical environment data to which the optical element to be processed is subjected during processing;

[0008] Inputting the thickness data of the optical element to be processed and the process parameter data into a preset optical element quality prediction model to obtain predicted thickness data of the optical element to be processed when processing in the optical element processing area is completed, wherein the optical element quality prediction model is obtained by optimizing and training a machine learning model based on a swarm intelligence optimization algorithm in advance, and is used to predict the thickness data of the optical element to be processed when processing in the optical element processing area is completed;

[0009] According to the predicted thickness data, it is determined whether quality problems will occur when processing the optical element to be processed in the optical element processing area.

[0010] In a possible design of the first aspect, the process of acquiring the optical element quality prediction model includes:

[0011] According to the historical thickness data of the optical element to be processed and the historical process parameter data of the optical element processing area, the swarm intelligence optimization algorithm is used to optimize the parameters of the initial optical element quality prediction model to obtain the optimal parameter configuration scheme;

[0012] According to the optimal parameter configuration scheme, the parameters in the initial optical element quality prediction model are configured to obtain the optical element quality prediction model.

[0013] In a possible design of the first aspect, the method of optimizing the parameters of the initial optical element quality prediction model using the swarm intelligence optimization algorithm according to the historical thickness data of the optical element to be processed and the historical process parameter data of the optical element processing area to obtain the optimal parameter configuration scheme includes:

[0014] Step a, acquiring a sample data set based on the historical thickness data of the optical element to be processed and the historical process parameter data of the optical element processing area, wherein the sample data set includes multiple groups of data samples, each group of data samples includes first historical thickness data of the optical element to be processed before processing, second historical thickness data of the optical element to be processed when processing is completed in the optical element processing area, and the historical process parameter data, wherein the historical process parameter data includes historical temperature data, historical humidity data, historical pressure data of the optical element processing area, and historical pressure data and historical optical environment data of the optical element to be processed during processing;

[0015] Step b, initializing I groups of population individuals, adaptive update quantity, adaptive update radius, adaptive mutation probability, adaptive penalty parameter factor, adaptive kernel parameter factor and preset iteration termination condition, wherein I is a positive integer, each group of population individuals represents a group of first parameter configuration schemes, and each group of first parameter configuration schemes includes a penalty parameter and a kernel parameter;

[0016] Step c: constructing a prediction model with exactly the same structure based on the machine learning model to obtain a first initial optical element quality prediction model;

[0017] Step d: for each first initial optical element quality prediction model, select a set of first parameter configuration schemes, configure the initial parameters of the first initial optical element quality prediction model, and obtain I second initial optical element quality prediction model, wherein the initial parameters of the first initial optical element quality prediction model include an initial penalty parameter and an initial kernel parameter;

[0018] Step e, selecting I groups of sample data from the sample data set, and inputting the first historical thickness data and the historical process parameter data in the I groups of sample data into I second initial optical element quality prediction models for prediction, to obtain I predicted thickness data;

[0019] Step f, calculating the fitness value of each group of first parameter configuration schemes according to the I predicted thickness data, the second historical thickness data in the I group of sample data, and the fitness function of the swarm intelligence optimization algorithm;

[0020] Step g, obtaining the first parameter configuration scheme corresponding to the highest fitness value, and recording it as the target parameter configuration scheme;

[0021] Step h: If the preset iteration termination condition is not met, the first parameter configuration scheme is updated according to the fitness value of each group of first parameter configuration schemes, the adaptive update quantity, the adaptive update radius, the adaptive mutation probability, the adaptive penalty parameter factor and the adaptive kernel parameter factor to obtain I groups of new first parameter configuration schemes, and steps d to h are repeated until the preset iteration termination condition is met, and the target parameter configuration is determined as the optimal parameter configuration scheme.

[0022] In a possible design of the first aspect, the first parameter configuration scheme is updated according to the fitness value of each group of first parameter configuration schemes, the adaptive update number, the adaptive update radius, the adaptive mutation probability, the adaptive penalty parameter factor, and the adaptive kernel parameter factor to obtain I groups of new first parameter configuration schemes, including:

[0023] Calculate the update quantity of each group of first parameter configuration schemes according to the fitness value of each group of first parameter configuration schemes and the adaptive update quantity;

[0024] Calculating the update radius of each group of first parameter configuration schemes according to the fitness value of each group of first parameter configuration schemes and the adaptive update radius;

[0025] According to the update quantity and the update radius, each group of first parameter configuration schemes is updated to obtain a second parameter configuration scheme;

[0026] randomly selecting a preset number of parameter configuration schemes from the second parameter configuration schemes as third parameter configuration schemes, and performing a mutation operation on the third parameter configuration scheme according to the adaptive mutation probability to obtain a fourth parameter configuration scheme;

[0027] According to the adaptive penalty parameter factor, adjusting the penalty parameter in the second parameter configuration scheme and the penalty parameter in the fourth parameter configuration scheme, and according to the adaptive core parameter factor, adjusting the core parameter in the second parameter configuration scheme and the core parameter in the fourth parameter configuration scheme to obtain adjusted second parameter configuration scheme and fourth parameter configuration scheme;

[0028] According to a preset selection strategy, a group I of new first parameter configuration schemes is selected from the adjusted second parameter configuration schemes and the fourth parameter configuration schemes.

[0029] In a possible design of the first aspect, determining, based on the predicted thickness data, whether quality problems will occur when processing the optical element to be processed in the optical element processing area includes:

[0030] If the predicted thickness data is not within the preset thickness data range, it is determined that quality problems will occur when processing the optical element to be processed in the optical element processing area.

[0031] In a possible design of the first aspect, the method further includes:

[0032] If it is determined that quality problems will occur when processing the optical element to be processed in the optical element processing area, the process parameter data of the optical element processing area is automatically adjusted, and quality warning information is pushed to the terminal device of the administrator, wherein the quality warning information is used to indicate that quality problems will occur when processing the optical element to be processed in the optical element processing area.

[0033] In a possible design of the first aspect, the adaptive update quantity is calculated using the following formula, including:

[0034]

[0035] in, Indicates The iteration The number of adaptive updates corresponding to the first parameter configuration scheme of the group, Indicates the preset maximum number of updates. Indicates the preset minimum number of updates. Indicates the preset maximum number of iterations, Indicates the current iteration number, Indicates the tth iteration The fitness value corresponding to the first parameter configuration scheme of the group, represents the maximum fitness value at the tth iteration, Indicates a preset constant;

[0036] Accordingly, the adaptive update radius is calculated using the following formula, including:

[0037]

[0038] in, Indicates The iteration The adaptive update radius corresponding to the first parameter configuration scheme of the group, Indicates the preset maximum update radius. Indicates the preset minimum update radius, Represents the minimum fitness value at the tth iteration.

[0039] In a second aspect, the present application provides an optical element production data processing device, comprising:

[0040] An acquisition module, used to acquire thickness data of the optical element to be processed and process parameter data of the optical element processing area based on the data uploaded by the sensor device, wherein the process parameter data includes temperature data, humidity data, pressure data of the optical element processing area, and pressure data and optical environment data to which the optical element to be processed is subjected during processing;

[0041] a processing module, for inputting the thickness data of the optical element to be processed and the process parameter data into a preset optical element quality prediction model, to obtain predicted thickness data of the optical element to be processed when processing in the optical element processing area is completed, wherein the optical element quality prediction model is obtained by optimizing and training a machine learning model based on a swarm intelligence optimization algorithm in advance, and is a model for predicting thickness data of the optical element to be processed when processing in the optical element processing area is completed;

[0042] A determination module is used to determine whether quality problems will occur when processing the optical element to be processed in the optical element processing area based on the predicted thickness data.

[0043] In a third aspect, the present application provides an electronic device, comprising: a receiver, a processor, and a memory connected to the processor;

[0044] The memory stores computer-executable instructions;

[0045] The processor executes the computer-executable instructions stored in the memory to implement the optical element production data processing method as described in any one of the first aspects.

[0046] In a fourth aspect, the present application provides an optical element production data processing system, comprising: the electronic device described in the third aspect and a sensor device connected to the electronic device;

[0047] The sensor device is used to collect thickness data of the optical element to be processed and process parameter data of the optical element processing area in real time and upload the collected data to the electronic device;

[0048] The sensor equipment includes: a spectral confocal sensor, a temperature sensor, a humidity sensor, a pressure sensor, a pressure sensor, and a fiber wave sensor;

[0049] The spectral confocal sensor is used to collect thickness data of the optical element to be processed;

[0050] The temperature sensor is used to collect temperature data of the optical element processing area;

[0051] The humidity sensor is used to collect humidity data of the optical element processing area;

[0052] The pressure sensor is used to collect pressure data of the optical element processing area;

[0053] The pressure sensor is used to collect pressure data on the optical element to be processed during processing;

[0054] The fiber optic wave sensor is used to collect optical environment data of the optical element to be processed during processing.

[0055] In a fifth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer execution instructions, which, when executed by a processor, are used to implement the optical element production data processing method as described in any one of the first aspects.

[0056] In a sixth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the optical element production data processing method as described in any one of the first aspects.

[0057] The optical element production data processing method and system provided in the present application relate to optical element detection technology. In the optical element processing area, the sensor equipment obtains the thickness data of the optical element to be processed and the process parameter data of the optical element processing area in real time, and transmits it to the electronic equipment to ensure the timeliness and accuracy of the detection. The electronic equipment obtains the predicted thickness data of the optical element to be processed when the processing in the optical element processing area is completed based on the preset optical element quality prediction model and the real-time transmitted data, and judges whether there will be quality problems based on this. This method can ensure the reliability and stability of the processing process, significantly improve the quality of optical elements, and effectively solve the limitations faced by the existing technical solutions in improving the quality of optical elements. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0059] Figure 1 A schematic diagram of an application scenario of an optical element production data processing method provided in this application;

[0060] Figure 2 A schematic diagram of a process flow of a first embodiment of a method for processing optical element production data provided by the present application;

[0061] Figure 3 A schematic diagram of a flow chart of a second embodiment of a method for processing optical element production data provided by the present application;

[0062] Figure 4 A schematic diagram of a process flow of a third embodiment of a method for processing optical element production data provided in the present application;

[0063] Figure 5 A schematic diagram of a flow chart of a fourth embodiment of a method for processing optical element production data provided in the present application;

[0064] Figure 6 A schematic diagram of a process flow of a fifth embodiment of a method for processing optical element production data provided by the present application;

[0065] Figure 7 A schematic diagram of a flow chart of a sixth embodiment of a method for processing optical element production data provided in the present application;

[0066] Figure 8 A schematic diagram of the structure of a first embodiment of an optical element production data processing device provided by the present application;

[0067] Fig. 9 A schematic diagram of the structure of an electronic device provided in this application;

[0068] Fig.10 A schematic diagram of the structure of an optical element production data processing system provided in this application.

[0069] The above drawings have shown clear embodiments of the present application, which will be described in more detail later. These drawings and text descriptions are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0070] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0071] With the rapid development of optical component inspection technology, various new and high-performance optical components have shown broad application prospects in emerging technology fields such as 3D imaging, 3D vision, and augmented reality. However, in the manufacturing process of optical components, quality inspection is a key link to ensure component performance.

[0072] In the existing technical solutions, optical component quality processing technologies mainly include cleaning, grinding, polishing and coating technologies.

[0073] However, existing optical component quality processing technologies still have certain limitations in improving the quality of optical components. First, cleaning technology has limited efficiency in removing tiny particles and nano-scale contaminants, which may cause residues to affect optical performance. Secondly, although grinding and polishing technologies can improve surface finish, they are prone to introduce new surface defects or cause material loss when processing micro-nano-scale components. Finally, in the process of achieving a uniform and dense thin film layer, coating technology may face problems such as insufficient film adhesion and uneven thickness, which will affect the overall performance and service life of the component. Therefore, the development of more efficient, precise and reliable quality processing technology is still an urgent need to improve the quality of optical components.

[0074] In response to the above problems, the present application discovered during the research on the optical element production data processing method that the existing optical element quality processing technology requires superb process level and fine operating skills, and the operation is difficult. At the same time, the existing optical element quality processing technology still has certain limitations in improving the quality of optical elements. Based on this, the present application considers whether it is possible to improve the quality of optical elements from the perspective of prediction algorithms. Specifically, in the optical element processing area, the sensor equipment obtains the thickness data of the optical element to be processed and the process parameter data of the optical element processing area in real time, and transmits it to the optical element quality prediction model in the electronic device to obtain the predicted thickness data of the optical element to be processed when the processing in the optical element processing area is completed. Based on the predicted thickness data, the possibility of quality problems is judged, thereby improving the quality of optical elements to a certain extent.

[0075] Figure 1 A schematic diagram of an application scenario of an optical element production data processing method provided in this application. Figure 1As shown, the optical element production data processing method provided by the present application includes a sensor device 101 and an electronic device 102. Among them, the sensor device 101 is used to obtain the thickness data of the optical element to be processed and the process parameter data of the optical element processing area. The optical element quality prediction model 1021 pre-deployed in the electronic device 102 is mainly used to process the thickness data of the optical element to be processed and the process parameter data of the optical element processing area uploaded by the sensor device 101, and to determine whether the optical element to be processed will have quality problems when processed in the optical element processing area.

[0076] Sensor device 101 includes but is not limited to spectral confocal sensors, temperature sensors, humidity sensors, pressure sensors, pressure sensors, fiber wave sensors, etc. Electronic devices 102 include but are not limited to servers, personal computers, laptops, tablet computers, and smart phones. These devices are capable of running complex algorithms and models, processing real-time data from sensors, and providing efficient computing power and storage space to support the operation of optical component quality prediction models. Figure 1 Only one sensor device 101 and electronic device 102 are shown in FIG. 1 , but it should be understood that two or more sensor devices 101 and electronic devices 102 may exist.

[0077] Specifically, in the optical element processing area, the sensor device 101 obtains the thickness data of the optical element to be processed and the process parameter data of the optical element processing area in real time, and transmits the thickness data of the optical element to be processed and the process parameter data of the optical element processing area to the electronic device 102 through the mobile network and / or the wireless network. The electronic device 102 processes the received thickness data of the optical element to be processed and the process parameter data of the optical element processing area based on the internally deployed optical element quality prediction model 1021, and obtains the predicted thickness data of the optical element to be processed when the processing in the optical element processing area is completed, so as to determine whether there will be quality problems when the optical element to be processed is processed in the optical element processing area.

[0078] It should be noted that the above scenario is only an example of an application scenario provided by an embodiment of the present application. The embodiment of the present application does not limit the actual form of the various devices included in the scenario. In the specific application of the solution, it can be set according to actual needs.

[0079] For ease of description, the following describes the implementation of the method by taking the execution subject as an electronic device capable of executing the optical element production data processing method as an example. It can be understood that the execution subject of the method as an electronic device is only an exemplary description and should not be understood as a limitation of the method.

[0080] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0081] Embodiment 1

[0082] Figure 2 This is a flow chart of a first embodiment of a method for processing optical element production data provided by the present application. Figure 2 As shown, the process of the optical element production data processing method may include:

[0083] S201: Based on the data uploaded by the sensor device, the thickness data of the optical element to be processed and the process parameter data of the optical element processing area are obtained, and the process parameter data include the temperature data, humidity data, pressure data of the optical element processing area, and the pressure data and optical environment data to which the optical element to be processed is subjected during processing.

[0084] In this step, in the manufacturing process of optical elements, thickness is a key quality indicator, which directly affects the performance and application effect of optical elements. However, due to the multiple variables and complex process parameters in the processing process, the thickness of optical elements is prone to deviation, leading to quality problems. Among them, the real-time acquisition and accurate measurement of the thickness data and process parameter data of the optical elements to be processed can provide detailed processing information, help identify and predict potential quality problems. Therefore, in order to improve the quality of optical elements, it is crucial to obtain the thickness data of the optical elements to be processed and the process parameter data of the processing area of ​​the optical elements. It is worth noting that optical elements include but are not limited to any one or any combination of diffractive optical elements, micro-nano optical elements, reconfigurable micro-nano optical elements based on phase change, new optical films, and optical metasurfaces.

[0085] Preferably, the optical element is a micro-nano optical element. Specifically, due to its micron or nanometer structure and characteristic size, the micro-nano optical element has the advantages of small size, thin thickness and light weight, so that it can replace traditional optical elements, thereby promoting the miniaturization and integration of optical systems. At the same time, micro-nano optical elements can modulate more complex light fields, and therefore show wide application potential in emerging technology fields such as three-dimensional imaging, three-dimensional vision and augmented reality. However, due to the extremely small size of micro-nano optical elements, thickness detection before and after processing has become a key technical problem at present. Therefore, accurately measuring the thickness of micro-nano optical elements is crucial to judging the processing effect and ensuring the quality of the elements.

[0086] Specifically, in order to timely obtain the thickness data of the optical element to be processed and the process parameter data of the optical element processing area, the present application pre-installs at least one sensor device in the optical element processing area, and the sensor device includes but is not limited to any one or any combination of spectral confocal sensors, temperature sensors, humidity sensors, pressure sensors, pressure sensors, fiber wave sensors, etc. The processing area refers to the specific environment and position where the optical element is located during the manufacturing process, including but not limited to the areas where the processing equipment, workbenches, control systems, etc. are located. The real-time data obtained by the perception layer device includes but is not limited to the thickness data of the optical element to be processed, the temperature data, humidity data, pressure data of the optical element processing area, and any one or any combination of the pressure data and optical environment data to which the optical element to be processed is subjected during processing.

[0087] Among them, the thickness data of the optical element to be processed is used to compare with the predicted thickness data after processing, which directly reflects the processing accuracy of the optical element to determine whether there will be quality problems. Temperature data and humidity data provide thermodynamic conditions in the processing area of ​​the optical element, which will affect the physical properties of the material and the processing effect. Pressure data and pressure data reflect the mechanical stress applied during the processing, which will affect the shape and surface quality of the optical element. Optical environment data includes information such as light intensity and spectral distribution, which has a direct impact on the final performance of the optical element. By comprehensively analyzing these real-time data, we can fully understand the various influencing factors in the processing process, thereby optimizing process parameters, preventing potential quality problems, and ultimately improving the overall quality and performance of optical components.

[0088] After the sensor device obtains the thickness data and process parameter data of the optical element to be processed, it will promptly transmit these data to the electronic device through the mobile network and / or wireless network for prediction and analysis of the thickness data of the optical element to be processed when the processing is completed.

[0089] Optionally, the data transmission method between the sensor device and the electronic device includes but is not limited to any one or any combination of direct connection method, gateway transmission method, and wireless communication method. Specifically, the direct connection method means that if the sensor device has a network interface (such as Ethernet, Wi-Fi) function and can be connected to the Internet, the network parameters such as Internet Protocol address, subnet mask, gateway and domain name system are set on the sensor device to ensure that the sensor device can access the network where the electronic device is located. The sensor device transmits the acquired data directly to the electronic device through the transmission control protocol / Internet protocol protocol. The gateway transmission method refers to connecting the sensor device to a local area network gateway (such as a router, switch, etc.), and the gateway is responsible for forwarding the data of the sensor device to the electronic device, wherein the gateway can use a preset protocol to transmit data with the electronic device, and the gateway is connected to the Internet through Ethernet, Wi-Fi, cellular network, etc., and transmits the data acquired by the sensor device to the electronic device. Wireless communication means that if the distance between the sensor device and the electronic device is far or the transmission method is limited, wireless communication technology can be used to transmit data to an intermediate device (such as a mobile phone, embedded system), and the intermediate device is connected to the Internet, and the corresponding communication protocol and parameters are set. The sensor device transmits the data to the intermediate device wirelessly, and the intermediate device then uploads the data to the electronic device. Preferably, the electronic device is a cloud computer.

[0090] Optionally, the application scenarios of the present application include but are not limited to any one or any combination of the following: on the production line of an optical component manufacturing plant, for real-time monitoring of the quality of optical components; in scientific research institutions, for research and development testing of new optical components to improve their quality; in university laboratories, for thickness verification and quality inspection of optical component samples.

[0091] S202: Input the thickness data and process parameter data of the optical element to be processed into a preset optical element quality prediction model to obtain the predicted thickness data of the optical element to be processed when processing in the optical element processing area is completed. The optical element quality prediction model is obtained by optimizing and training the machine learning model based on the swarm intelligence optimization algorithm in advance, and is a model for predicting the thickness data of the optical element to be processed when processing in the optical element processing area is completed.

[0092] In this step, based on step S201, after receiving the thickness data and process parameter data of the optical element to be processed uploaded by the sensor device, the electronic device inputs these data into a preset optical element quality prediction model for prediction processing to obtain the predicted thickness data of the optical element to be processed when the processing in the optical element processing area is completed.

[0093] Among them, the preset optical component quality prediction model is obtained by pre-optimizing and training the machine learning model based on the swarm intelligence optimization algorithm. It is a model used to predict the thickness data of the optical component to be processed when processing is completed in the optical component processing area. The model analyzes the historical thickness data and historical process parameter data of the optical component to be processed, combined with the thickness data and process parameter data of the optical component to be processed uploaded in real time by the perception layer equipment, and can effectively predict the thickness data of the optical component to be processed when processing is completed in the optical element processing area.

[0094] Specifically, swarm intelligence optimization algorithms are a type of optimization algorithm that simulates group behavior in nature and seeks the optimal solution through collaboration and information sharing between individuals. Swarm intelligence optimization algorithms include, but are not limited to, any one or any combination of particle swarm optimization algorithms, ant colony algorithms, artificial bee colony algorithms, and enhanced fireworks algorithms. By applying swarm intelligence optimization algorithms, we can effectively search complex multidimensional spaces and find the global optimal solution, thereby improving the prediction accuracy and robustness of machine learning models.

[0095] Machine learning models are a technology that uses computing technology to build prediction models through a data-driven approach. Specifically, machine learning models can learn patterns from a large amount of historical data and apply them to the prediction of new data. By combining swarm intelligence optimization algorithms, machine learning models can better adapt to the relationship between complex process parameters and thickness data, thereby improving the accuracy of predictions.

[0096] S203: Determine whether quality problems will occur when processing the optical element to be processed in the optical element processing area based on the predicted thickness data.

[0097] In this step, based on step S202, after the electronic device obtains the predicted thickness data of the optical element to be processed when the processing in the optical element processing area is completed, it determines whether there will be quality problems in the processing of the optical element to be processed in the optical element processing area according to the predicted thickness data.

[0098] Specifically, by comparing the predicted thickness data with the thickness data of the optical element to be processed obtained by the sensor device, it can be accurately determined whether quality problems will occur when processing the optical element to be processed under the current process parameters. Through this comparison, the optical element quality prediction model can identify any deviations and determine whether these deviations exceed the preset quality range. If the deviation is found to exceed the preset quality range, the electronic equipment that deploys the optical element quality prediction model can promptly issue an alarm and recommend adjusting the process parameters to prevent potential quality problems.

[0099] It is worth noting that the specific implementation of this step can refer to step S601 in the subsequent embodiment 5, which will not be repeated here.

[0100] The optical element production data processing method provided in the embodiment of the present application is based on the optical element quality prediction model, and processes the thickness data and process parameter data of the optical element to be processed acquired by the sensor device in real time, so as to obtain the predicted thickness data of the optical element to be processed when the processing is completed in the optical element processing area, so as to judge whether there will be quality problems. By making full use of the thickness data and process parameter data acquired in real time, the method can timely discover and prevent potential quality problems, thereby improving the performance, production efficiency and reliability of the optical element.

[0101] Next, the process of obtaining the optical component quality prediction model will be discussed in detail.

[0102] Embodiment 2

[0103] Figure 3 This is a flow chart of Embodiment 2 of a method for processing optical element production data provided by the present application. Figure 3 As shown, the process of the optical element production data processing method may include:

[0104] S301: According to the historical thickness data of the optical element to be processed and the historical process parameter data of the optical element processing area, a swarm intelligence optimization algorithm is used to optimize the parameters of the initial optical element quality prediction model to obtain an optimal parameter configuration solution.

[0105] In this step, when applying the preset optical quality prediction model to obtain the predicted thickness data of the optical element to be processed when the processing in the optical element processing area is completed, it is necessary to obtain the preset optical quality prediction model in advance.

[0106] Specifically, we first need to collect and organize the historical thickness data of the optical components to be processed and the historical process parameter data of the optical component processing area. These data include the thickness changes of the optical components and the corresponding process parameters under different processing conditions. Through in-depth analysis of these historical data, the key factors affecting the thickness changes of optical components can be identified. Next, the swarm intelligence optimization algorithm is used to optimize the parameters in the initial optical component quality prediction model. During the optimization process, the swarm intelligence optimization algorithm will repeatedly adjust the parameters of the initial optical component quality prediction model to minimize the prediction error or maximize the prediction accuracy.

[0107] In one possible implementation, the swarm intelligence optimization algorithm first generates a set of initial parameter configurations and trains the initial optical element quality prediction model based on these parameter configurations. Then, the swarm intelligence optimization algorithm adjusts the parameter configurations based on the prediction results and gradually approaches the optimal solution. Through multiple iterations, the swarm intelligence optimization algorithm will eventually converge to an optimal parameter configuration scheme, so that the initial optical element quality prediction model can more accurately predict the thickness change of the optical element to be processed, thereby improving the accuracy and reliability of quality control.

[0108] It is worth noting that the specific implementation of this step can refer to steps a to h in the subsequent embodiment 3, and will not be repeated here.

[0109] S302: According to the optimal parameter configuration scheme, configure the parameters in the initial optical component quality prediction model to obtain the optical component quality prediction model.

[0110] In this step, according to step S301, after obtaining the optimal parameter configuration scheme, the parameters in the initial optical element quality prediction model are configured according to the optimal parameter configuration scheme, thereby obtaining the optical element quality prediction model.

[0111] Specifically, according to the network structure and optimal parameter configuration scheme of the predefined initial optical element quality prediction model, the parameters of the corresponding positions in the initial optical element quality prediction model are determined, so as to obtain an optical element quality prediction model for predicting the thickness data of the optical element to be processed when processing is completed in the optical element processing area.

[0112] The optical element production data processing method provided in the embodiment of the present application mainly involves the process of obtaining the optical element quality prediction model. Among them, the parameters in the optical element quality prediction model are determined according to the historical thickness data of the optical element to be processed, the historical process parameter data of the optical element processing area, and the swarm intelligence optimization algorithm. Through this method, the optimal parameter configuration scheme in the optical element quality prediction model can be accurately determined, and the prediction accuracy and reliability of the optical element quality prediction model can be improved.

[0113] Next, the specific implementation process of step S301 will be discussed in detail.

[0114] Embodiment 3

[0115] Figure 4 This is a flow chart of Embodiment 3 of a method for processing optical element production data provided by the present application. Figure 4As shown, based on any of the above embodiments, according to the historical thickness data of the optical element to be processed and the historical process parameter data of the optical element processing area, a swarm intelligence optimization algorithm is used to optimize the parameters of the initial optical element quality prediction model to obtain the optimal parameter configuration scheme. The process of the optical element production data processing method may include:

[0116] Step a, obtaining a sample data set based on historical thickness data of the optical element to be processed and historical process parameter data of the optical element processing area, the sample data set including multiple groups of data samples, each group of data samples including first historical thickness data of the optical element to be processed before processing, second historical thickness data of the optical element to be processed when processing is completed in the optical element processing area, and historical process parameter data, the historical process parameter data including historical temperature data, historical humidity data, historical pressure data of the optical element processing area, and historical pressure data and historical optical environment data of the optical element to be processed during processing.

[0117] In this step, when using the swarm intelligence optimization algorithm to determine the optimal parameters in the optical element quality prediction model, it is necessary to obtain in advance the historical thickness data of the optical element to be processed and the historical process parameter data of the optical element processing area to determine the sample data set.

[0118] The historical thickness data of the optical element to be processed includes the first historical thickness data of the optical element to be processed before processing and the second historical thickness data of the optical element to be processed after processing in the optical element processing area. The first historical thickness data indicates the thickness of the optical element to be processed before processing, which is used to understand the initial state of the optical element. The second historical thickness data indicates the thickness of the optical element to be processed after processing in the optical element processing area, which is used to determine the influence of the processing process on the thickness of the optical element.

[0119] The historical process parameter data of the optical element processing area include the historical temperature data, historical humidity data, historical pressure data of the optical element processing area, as well as the historical pressure data and historical optical environment data to which the optical element to be processed is subjected during processing. They are used to describe the processing environment and conditions and have an important impact on the accuracy of the initial optical element quality prediction model.

[0120] The sample data set includes multiple groups of data samples, each group of data samples includes first historical thickness data of the unprocessed optical element, second historical thickness data of the processed optical element after processing in the optical element processing area, and historical process parameter data.

[0121] Optionally, the sample data set can be obtained through the thickness data of the optical element to be processed and the process parameter data of the optical element processing area previously collected by the measuring equipment. Specifically, the technician first uses measuring equipment, such as sensor equipment, step meter and optical film thickness measuring instrument, to collect data on the prepared optical element samples, and record the process parameters and environmental conditions during the production process. It is worth noting that the measuring equipment is adjusted and selected according to the type of optical element and the specific circumstances of the processing technology. Then, the optical sample is processed using femtosecond laser direct writing technology or other suitable technologies. During the processing process, the technician needs to install and set the parameters of the sensor equipment in advance, and upload the collected sensor data to the electronic device through a suitable network interface.

[0122] Optionally, before determining the sample data set based on the historical thickness data of the optical element to be processed and the historical process parameter data of the optical element processing area, preprocessing operations can be performed on these data. For example, processing missing data, data standardization, data denoising, and outlier processing. Specifically, due to equipment failure, data transmission errors, or human operating errors, the historical thickness data of the optical element to be processed and the historical process parameter data of the optical element processing area may have missing values ​​or outliers. In order to improve the training effect and efficiency of the optical element quality prediction model, the standard deviation normalization method can be used to standardize the various feature columns of the historical thickness data of the optical element to be processed and the historical process parameter data of the optical element processing area. In addition, interpolation or mean filling methods can be used to process missing data, and filtering and other techniques can be used to remove noise and outliers in the data, thereby effectively improving the prediction accuracy and reliability of the initial optical element quality prediction model.

[0123] Step b, initialize I groups of population individuals, adaptive update quantity, adaptive update radius, adaptive mutation probability, adaptive penalty parameter factor, adaptive kernel parameter factor and preset iteration termination condition, wherein I is a positive integer, each group of population individuals represents a group of first parameter configuration schemes, and each group of first parameter configuration schemes includes a penalty parameter and a kernel parameter.

[0124] In this step, after obtaining multiple groups of data samples according to step a, the hyperparameters in the swarm intelligence optimization algorithm are initialized to optimize the parameters in the initial optical element quality prediction model.

[0125] Among them, the hyperparameters in the swarm intelligence optimization algorithm include but are not limited to I groups of swarm individuals, adaptive update number, adaptive update radius, adaptive mutation probability, adaptive penalty parameter factor, adaptive kernel parameter factor and preset iteration termination condition, and I is a positive integer. It is worth noting that I is determined according to actual conditions, and this application does not make specific restrictions on this.

[0126] Specifically, I groups of group individuals represent I groups of first parameter configuration schemes. In the swarm intelligence optimization algorithm, each group of group individuals represents a group of first parameter configuration schemes. Each group of first parameter configuration schemes includes a penalty parameter and a kernel parameter.

[0127] The penalty parameter is used to control the emphasis of the optical component quality prediction model on the constraints during the optimization process. In the optical component quality prediction model, the penalty parameter can help balance the complexity and fitting accuracy of the model to prevent overfitting or underfitting. For example, in the support vector machine model, the penalty parameter is used to control the penalty intensity of misclassified samples. A larger penalty parameter will make the model fit the training data more strictly, while a smaller penalty parameter will allow more misclassifications, thereby improving the generalization ability of the model.

[0128] The kernel parameter is used to control the characteristics of the kernel function and affect the nonlinear mapping ability of the optical component quality prediction model. In the optical component quality prediction model, the kernel parameter determines how the model maps the input data to a high-dimensional feature space in order to better capture the complex relationship of the data. For example, in the radial basis function kernel, the kernel parameter controls the width of the Gaussian distribution. A larger kernel parameter will make the Gaussian distribution narrower, so that the model is more inclined to fit local data, while a smaller kernel parameter will make the Gaussian distribution wider, so that the model is more inclined to fit global data.

[0129] By initializing the first parameter configuration scheme of group I, the swarm intelligence optimization algorithm can explore different penalty parameter and kernel parameter combinations in the search space to find the optimal parameter configuration scheme. These optimal parameter configuration schemes will be used to optimize the initial optical component quality prediction model to improve the prediction accuracy and reliability of the model.

[0130] The adaptive update number refers to the number of individuals that the swarm intelligence optimization algorithm adaptively updates based on the performance of the current group in each iteration. The adaptive update number can help the swarm intelligence optimization algorithm dynamically adjust the search intensity at different search stages and improve optimization efficiency.

[0131] Adaptive update radius means that in each iteration, the swarm intelligence optimization algorithm adaptively adjusts the search radius according to the performance of the current group. Adaptive update radius can help the swarm intelligence optimization algorithm to conduct extensive search in the early stage and fine search in the later stage, thereby improving the convergence speed and accuracy.

[0132] Adaptive mutation probability means that in each iteration, the swarm intelligence optimization algorithm adaptively adjusts the probability of individual mutation according to the performance of the current group. Adaptive mutation probability can help the swarm intelligence optimization algorithm dynamically adjust the mutation intensity at different search stages, avoid falling into local optimal solutions, and improve global search capabilities.

[0133] Adaptive penalty parameter factor means that in each iteration, the swarm intelligence optimization algorithm adaptively adjusts the penalty parameter factor according to the performance of the current group. The adaptive penalty parameter factor is used to control the penalty intensity of the constraint conditions, ensure that the swarm intelligence optimization algorithm meets the constraint conditions during the search process, and improve the feasibility of the solution.

[0134] Adaptive kernel parameter factors refer to the fact that in each iteration, the swarm intelligence optimization algorithm adaptively adjusts the kernel parameter factors according to the performance of the current group. The adaptive kernel parameter factors are used to control the parameters of the kernel function and affect the search range and accuracy of the swarm intelligence optimization algorithm.

[0135] The preset iteration termination condition refers to the preset stop condition during the execution of the swarm intelligence optimization algorithm. The preset iteration termination condition includes but is not limited to any one or any combination of the maximum number of iterations, the objective function value reaching the preset threshold, the change amplitude of the group individual less than the preset value, etc. The preset iteration termination condition is used to control the running time and convergence speed of the swarm intelligence optimization algorithm to ensure that the optimal solution is found within a reasonable time.

[0136] Step c: constructing I prediction models with exactly the same structure based on the machine learning model to obtain I first initial optical element quality prediction models.

[0137] In this step, in order to determine the optimal parameter configuration scheme in the initial optical element quality prediction model, it is necessary to construct I prediction models with exactly the same structure based on the machine learning model in advance, so as to obtain I first initial optical element quality prediction models.

[0138] It is worth noting that each first initial optical element quality prediction model is completely consistent in structure. By constructing a prediction model with exactly the same structure, parameter optimization can be performed under the same model structure to ensure the comparability and consistency of the optimization results.

[0139] Step d: for each first initial optical element quality prediction model, select a set of first parameter configuration schemes, configure the initial parameters of the first initial optical element quality prediction model, and obtain I second initial optical element quality prediction models, wherein the initial parameters of the first initial optical element quality prediction model include initial penalty parameters and initial kernel parameters.

[0140] In this step, after obtaining I first initial optical element quality prediction models based on step c, a set of first parameter configuration schemes is selected for each first initial optical element quality prediction model, and the initial parameters in the first initial optical element quality prediction model are configured to obtain I second initial optical element quality prediction models.

[0141] The initial parameters of the first initial optical element quality prediction model include initial penalty parameters and initial kernel parameters, and the number and meaning of these initial parameters are consistent with those in the first parameter configuration scheme.

[0142] Step e: select I groups of sample data from the sample data set, and input the first historical thickness data and historical process parameter data in the I groups of sample data into I second initial optical element quality prediction models for prediction, to obtain I predicted thickness data.

[0143] In this step, after obtaining I second initial optical element quality prediction models according to step d, I groups of sample data are selected from the sample data set obtained in step a, and the first historical thickness data and historical process parameter data in the I groups of sample data are respectively input into I second initial optical element quality prediction models for prediction, thereby obtaining I predicted thickness data, which can reflect the prediction ability of each second initial optical element quality prediction model for given input data.

[0144] Step f, calculating the fitness value of each group of first parameter configuration schemes according to I predicted thickness data, the second historical thickness data in I groups of sample data, and the fitness function of the swarm intelligence optimization algorithm.

[0145] In this step, after obtaining I predicted thickness data according to step e, the fitness value of each group of first parameter configuration schemes is calculated based on the I predicted thickness data, the second historical thickness data in the I group of sample data, and the fitness function of the swarm intelligence optimization algorithm.

[0146] Specifically, the fitness function of the swarm intelligence optimization algorithm is used to quantify the quality of each group of first parameter configuration schemes. Among them, the fitness function is a function used to judge the quality of each group of first parameter configuration schemes in the swarm intelligence optimization algorithm, and its fitness value is mainly determined by calculating the performance of the first parameter configuration scheme on a specific problem. The higher the fitness value, the better the performance of the first parameter configuration scheme on the current problem. It is worth noting that the setting of the fitness function helps the swarm intelligence optimization algorithm to search and identify the optimal parameter configuration scheme more efficiently, and can also improve the accuracy and quality of the optical component processing process and improve production efficiency.

[0147] Optionally, the fitness function of the present application can be defined as an error function between the predicted thickness data and the actual thickness data (ie, the second historical thickness data). Specifically, the fitness function is calculated using the following formula:

[0148]

[0149] in, Indicates The fitness value corresponding to the first parameter configuration scheme of the group, Indicates The mean absolute error value corresponding to the first parameter configuration scheme of the group, Indicates The root mean square error value corresponding to the first parameter configuration scheme of the group.

[0150] Specifically, The calculation formula is as follows:

[0151]

[0152] in, Indicates The mean absolute error value corresponding to the first parameter configuration scheme of the group, represents the number of the first parameter configuration schemes, Indicates The actual thickness data corresponding to the first parameter configuration scheme (i.e., the second historical thickness data) is obtained. Indicates The predicted thickness data corresponding to the first parameter configuration scheme of the group.

[0153] The calculation formula is as follows:

[0154]

[0155] in, Indicates The root mean square error value corresponding to the first parameter configuration scheme of the group, represents the number of the first parameter configuration schemes, Indicates The actual thickness data corresponding to the first parameter configuration scheme (i.e., the second historical thickness data) is obtained. Indicates The predicted thickness data corresponding to the first parameter configuration scheme of the group.

[0156] The fitness value is calculated by taking into account the difference between the predicted thickness data and the actual thickness data (i.e., the second historical thickness data). In this way, the performance of each group of first parameter configuration schemes can be comprehensively judged, thereby better realizing the search process of the swarm intelligence optimization algorithm to determine the optimal parameter configuration scheme.

[0157] It is worth noting that the fitness function of the swarm intelligence optimization algorithm can be adjusted according to specific needs, and this application does not make specific limitations on this.

[0158] Step g: Obtain the first parameter configuration scheme corresponding to the highest fitness value and record it as the target parameter configuration scheme.

[0159] In this step, after the fitness value of each group of first parameter configuration schemes is calculated according to step f, the fitness values ​​of the I groups of first parameter configuration schemes are sorted from high to low or from low to high, so as to find the first parameter configuration scheme with the highest fitness value, which is recorded as the target parameter configuration scheme.

[0160] Step h: If the preset iteration termination condition is not met, the first parameter configuration scheme is updated according to the fitness value, adaptive update quantity, adaptive update radius, adaptive mutation probability, adaptive penalty parameter factor and adaptive kernel parameter factor of each group of first parameter configuration schemes to obtain I groups of new first parameter configuration schemes. Steps d to h are repeated until the preset iteration termination condition is met, and the target parameter configuration is determined as the optimal parameter configuration scheme.

[0161] In this step, after obtaining the first parameter configuration scheme corresponding to the highest fitness value according to step g, it is determined whether the current stage meets the preset iteration termination condition.

[0162] If the current stage meets the preset iteration termination condition, the obtained target parameter configuration scheme is determined as the optimal parameter configuration scheme. The optimal parameter configuration scheme is the optimal solution obtained through iteration and optimization, which can be used to configure the parameters in the initial optical element quality prediction model to obtain the optical element quality prediction model.

[0163] If the preset iteration termination condition is not met in the current stage, the first parameter configuration scheme is updated according to the fitness value, adaptive update quantity, adaptive update radius, adaptive mutation probability, adaptive penalty parameter factor and adaptive kernel parameter factor of each group of first parameter configuration schemes to obtain I groups of new first parameter configuration schemes, and steps d to h are repeated until the preset iteration termination condition is met. The specific implementation method can refer to steps S501 to S506 in the subsequent embodiment 4, which will not be repeated here.

[0164] For example, the preset iteration termination condition is the preset maximum number of iterations, and the preset maximum number of iterations is 600. If the current number of iterations is 467, continue to perform steps d to h. If the current number of iterations is 600, the obtained target parameter configuration scheme is determined as the optimal parameter configuration scheme.

[0165] The optical component production data processing method provided in the embodiment of the present application effectively determines the parameter configuration scheme of the initial optical component quality prediction model through a swarm intelligence optimization algorithm. Specifically, the swarm intelligence optimization algorithm repeatedly updates and optimizes the parameter configuration scheme according to a preset iteration termination condition until the optimal parameter configuration scheme is found. Through this iterative optimization method, the swarm intelligence optimization algorithm can gradually converge to the optimal solution in the continuous iteration process, thereby laying a solid foundation for building a high-precision optical component quality prediction model. At the same time, the method can also improve the prediction accuracy and reliability of the optical component quality prediction model and the overall efficiency of optical component quality control.

[0166] Next, the specific implementation process of step h will be discussed in detail.

[0167] Embodiment 4

[0168] Figure 5 This is a flow chart of a fourth embodiment of a method for processing optical element production data provided by the present application. Figure 5 As shown, based on any of the above embodiments, according to the fitness value, adaptive update quantity, adaptive update radius, adaptive mutation probability, adaptive penalty parameter factor and adaptive kernel parameter factor of each group of first parameter configuration schemes, the first parameter configuration scheme is updated to obtain a group of new first parameter configuration schemes. The process of the optical element production data processing method may include:

[0169] S501: Calculate the update quantity of each group of first parameter configuration solutions according to the fitness value and the adaptive update quantity of each group of first parameter configuration solutions.

[0170] In this step, when the preset iteration termination condition is not met in the current stage, the first parameter configuration scheme needs to be updated according to the fitness value, adaptive update number, adaptive update radius, adaptive mutation probability, adaptive penalty parameter factor and adaptive kernel parameter factor of each group of first parameter configuration schemes.

[0171] When updating each set of first parameter configuration schemes, the update quantity of each set of first parameter configuration schemes needs to be determined in advance. The adaptive update quantity refers to the quantity of each set of first parameter configuration schemes that needs to be updated, which is dynamically adjusted according to the current fitness value during each iteration.

[0172] Specifically, the update quantity of each group of first parameter configuration solutions can be calculated according to the fitness value and the adaptive update quantity of each group of first parameter configuration solutions.

[0173] In a possible implementation, the adaptive update quantity is calculated using the following formula:

[0174]

[0175] in, Indicates The iteration The number of adaptive updates corresponding to the first parameter configuration scheme of the group, Indicates the preset maximum number of updates. Indicates the preset minimum number of updates. Indicates the preset maximum number of iterations, Indicates the current iteration number, Indicates the tth iteration The fitness value corresponding to the first parameter configuration scheme of the group, represents the maximum fitness value at the tth iteration, Indicates a preset constant.

[0176] The setting of the adaptive update number can dynamically adjust the update number according to the convergence of the current swarm intelligence optimization algorithm. While improving the optimization efficiency and enhancing the robustness of the swarm intelligence optimization algorithm, it also improves the flexibility of the swarm intelligence optimization algorithm, thus laying the foundation for improving the quality of optical components.

[0177] Optionally, to prevent the number of adaptive updates from being too large or too small, the following calculation formula is used to limit the number of adaptive updates:

[0178]

[0179] in, is the preset minimum number of adaptive updates, The preset maximum number of adaptive updates.

[0180] Limiting the number of adaptive updates can ensure that the swarm intelligence optimization algorithm maintains a balance when searching the parameter space, avoiding overfitting or waste of computing resources due to too many updates. Too many adaptive updates may make the model sensitive to the noise of the training data, thereby affecting its generalization ability on new data and causing a decrease in the prediction quality of optical components. On the other hand, if the number of adaptive updates is too small, the swarm intelligence optimization algorithm may not be able to fully explore the parameter space, resulting in failure to find the optimal solution, thus affecting the training results of the model. When limiting the number of adaptive updates, the swarm intelligence optimization algorithm can maintain control over the complexity of the model while ensuring a certain exploration depth and efficiency. This method ensures the stability of the model training process, enabling it to provide accurate thickness predictions when facing different process parameters. By reasonably setting the number of adaptive updates, the model training process can reduce unnecessary iterations, improve computing efficiency, and reduce the consumption of time and resources, thereby improving the overall quality of optical components and enhancing the reliability and applicability of the model in practical applications.

[0181] S502: Calculate the update radius of each group of first parameter configuration schemes according to the fitness value and the adaptive update radius of each group of first parameter configuration schemes.

[0182] In this step, when updating each set of first parameter configuration schemes, the update radius of each set of first parameter configuration schemes needs to be determined in advance. The adaptive update radius refers to the amplitude or range of the parameter update in the first parameter configuration scheme dynamically adjusted according to the current fitness value during each iteration.

[0183] In a possible implementation, the adaptive update radius is calculated using the following formula:

[0184]

[0185] in, Indicates The iteration The adaptive update radius corresponding to the first parameter configuration scheme of the group, Indicates the preset maximum update radius. Indicates the preset minimum update radius, Represents the minimum fitness value at the tth iteration.

[0186] The setting of the adaptive update radius can dynamically adjust the search radius according to the convergence of the current swarm intelligence optimization algorithm, thereby preventing the swarm intelligence optimization algorithm from falling into the local optimal solution and enhancing the global search capability, thus significantly improving the overall performance and effect of the swarm intelligence optimization algorithm.

[0187] S503: Update each group of first parameter configuration schemes according to the update quantity and the update radius to obtain a second parameter configuration scheme.

[0188] In this step, after the update quantity calculated in step S501 and the update radius calculated in step S502 are obtained, each group of first parameter configuration schemes is updated to obtain a second parameter configuration scheme.

[0189] Specifically, the number of new parameter configuration schemes to be generated for each group of first parameter configuration schemes is determined according to the update number calculated in step S501. The search range of each group of first parameter configuration schemes is determined according to the update radius calculated in step S502.

[0190] Optionally, when updating each group of first parameter configuration schemes, the parameters in each group of first parameter configuration schemes may be randomly perturbed. The amplitude of the random perturbation is within the update radius, which means that the change in each parameter will not exceed the predefined update radius. In this way, a new parameter configuration scheme, i.e., a second parameter configuration scheme, may be generated based on the original parameter configuration scheme.

[0191] S504: Randomly select a preset number of parameter configuration schemes from the second parameter configuration scheme as a third parameter configuration scheme, and perform a mutation operation on the third parameter configuration scheme according to an adaptive mutation probability to obtain a fourth parameter configuration scheme.

[0192] In this step, in order to increase the diversity of parameter configuration schemes, after obtaining the second parameter configuration scheme according to step S503, a preset number of parameter configuration schemes are first randomly selected from the second parameter configuration scheme as the third parameter configuration scheme. It is worth noting that the preset number can be set according to specific needs and optimization goals, and this application does not make specific restrictions on this.

[0193] Next, the third parameter configuration scheme is mutated according to the adaptive mutation probability. The adaptive mutation probability is a mutation probability that is dynamically adjusted according to the current optimization state. Specifically, the adaptive mutation probability can be adjusted according to the current optimization state, the diversity of parameter configuration schemes, and other relevant factors. For example, in the early stage of optimization, a higher mutation probability can be set to increase the breadth of exploration. In the later stage of optimization, the mutation probability can be reduced to focus on local optimization.

[0194] By mutating the third parameter configuration scheme, new parameter values ​​can be introduced to further explore the parameter space and avoid falling into a local optimal solution. The mutation operation can be performed in a variety of ways, including but not limited to any one of Gaussian mutation, uniform mutation, etc. or any combination thereof. The specific mutation method can be adjusted according to the actual situation and optimization requirements, and this application does not make specific restrictions on this.

[0195] After the mutation operation, the fourth parameter configuration scheme can be obtained.

[0196] S505: According to the adaptive penalty parameter factor, the penalty parameter in the second parameter configuration scheme and the penalty parameter in the fourth parameter configuration scheme are adjusted, and according to the adaptive core parameter factor, the core parameter in the second parameter configuration scheme and the core parameter in the fourth parameter configuration scheme are adjusted to obtain the adjusted second parameter configuration scheme and the fourth parameter configuration scheme.

[0197] In this step, in order to prevent the parameters in the second parameter configuration scheme obtained in step S503 and the parameters in the fourth parameter configuration scheme obtained in step S504 from exceeding the preset parameter range, it is necessary to appropriately adjust the parameters in the second parameter configuration scheme and the fourth parameter configuration scheme that exceed the preset parameter range.

[0198] Specifically, before adjusting the parameters in the second parameter configuration scheme and the fourth parameter configuration scheme, it is necessary to predetermine the parameter configuration schemes corresponding to the parameters that exceed the preset parameter range. Assume that the number of parameter configuration schemes corresponding to the parameters that exceed the preset parameter range is , then according to the adaptive penalty parameter factor and the adaptive kernel parameter factor, adjust Penalty parameters and kernel functions in the group parameter configuration scheme.

[0199] Optionally, the penalty parameter in each parameter configuration scheme can be adjusted using the following formula:

[0200]

[0201] in, Indicates The iteration The adjusted penalty parameter in the group parameter configuration scheme, Indicates The iteration The penalty parameters not adjusted in the group parameter configuration scheme, represents the preset maximum penalty parameter, represents the preset minimum penalty parameter, represents a random function, Indicates The adaptive penalty parameter factor at the iteration.

[0202] Among them, The adaptive penalty parameter factor at the iteration The calculation is done using the following formula:

[0203]

[0204] in, No. The adaptive penalty parameter factor at the iteration, represents the preset maximum penalty parameter, Represents the preset minimum penalty parameter.

[0205] Optionally, the kernel parameters in each parameter configuration scheme can be adjusted using the following formula:

[0206]

[0207] in, Indicates The iteration The kernel parameters adjusted in the group parameter configuration scheme, Indicates The iteration The kernel parameters that are not adjusted in the group parameter configuration scheme, represents the preset maximum kernel parameter, represents the preset minimum kernel parameters, represents a random function, Indicates The adaptive kernel parameter factor at the iteration.

[0208] Among them, The adaptive kernel parameter factor at the iteration The calculation is done using the following formula:

[0209]

[0210] in, No. The adaptive kernel parameter factor at the iteration, represents the preset maximum kernel parameter, Indicates the preset minimum kernel parameters.

[0211] By adjusting the parameters in the second parameter configuration scheme and the fourth parameter configuration scheme that exceed the preset parameter range through the above calculation formula, it can be ensured that the parameters in the second parameter configuration scheme and the fourth parameter configuration scheme are within the preset parameter range, avoiding the generation of invalid parameter configuration schemes, thereby improving the stability and search efficiency of the swarm intelligence optimization algorithm.

[0212] S506: According to a preset selection strategy, select I group of new first parameter configuration schemes from the adjusted second parameter configuration scheme and the fourth parameter configuration scheme.

[0213] In this step, after obtaining the adjusted second parameter configuration scheme and the fourth configuration scheme according to step S505, a group of new first parameter configuration schemes is selected from the adjusted second parameter configuration scheme and the fourth parameter configuration scheme according to a preset selection strategy.

[0214] Optionally, the preset selection strategy can be determined according to the first preset fitness condition and the second preset fitness condition. Specifically, first, from all the adjusted second parameter configuration schemes and fourth parameter configuration schemes, the parameter configuration scheme whose fitness value meets the first preset fitness condition is determined as the first part of the first parameter configuration scheme, and then the parameter configuration scheme that meets the second preset fitness condition is determined as the second part of the first parameter configuration scheme.

[0215] It is worth noting that the I group of new first parameter configuration schemes includes a first part first parameter configuration scheme and a second part first parameter configuration scheme.

[0216] In a possible implementation, the first preset fitness condition is mainly set based on a reference factor in a ranking selection method, a threshold selection method, a percentile selection method, or any combination of reference factors.

[0217] Specifically, the sorting selection method first sorts all the adjusted second parameter configuration schemes and fourth parameter configuration schemes from high to low according to the fitness value, and then selects the parameter configuration schemes with the first preset number as the first parameter configuration scheme of the first part. This preset fitness condition can ensure that the selected parameter configuration scheme has a higher ranking in the fitness value, so that it performs better in practical applications.

[0218] The threshold selection rule sets a preset fitness value as a threshold and selects a parameter configuration scheme whose fitness value is greater than or equal to the preset fitness value as the first parameter configuration scheme of the first part. This preset fitness condition can ensure that the selected parameter configuration scheme meets a certain standard in fitness value, thereby ensuring its quality.

[0219] The percentile selection rule calculates the percentiles of all parameter configuration schemes based on the fitness values, and selects the parameter configuration schemes with fitness values ​​in the top several percentiles as the first parameter configuration scheme in the first part. For example, you can select the scheme with fitness values ​​in the top 20%. This method ensures that the selected scheme is at a higher position in the fitness value distribution, so it has better performance.

[0220] Through different first preset fitness conditions, the first parameter configuration scheme of the first part can be flexibly selected to ensure its quality and performance in practical applications. It is worth noting that the first preset fitness condition can be adjusted according to specific application requirements, and this application does not make specific limitations on this.

[0221] After the first part of the first parameter configuration scheme is determined to be completed, the second part of the first parameter configuration scheme is determined from all adjusted second parameter configuration schemes and fourth parameter configuration schemes according to the second preset fitness condition.

[0222] The second preset fitness condition is mainly set based on one reference factor or any combination of reference factors among the roulette method, the tournament selection method and the random sampling method.

[0223] Among them, if the second preset condition is the roulette method, the selection probability is first allocated according to the fitness value of each parameter configuration scheme, and then the roulette method is used for random selection to determine the first parameter configuration scheme of the second part. This method ensures that the parameter configuration scheme with high fitness has a higher selection probability, and also gives other parameter configuration schemes a certain selection opportunity, thereby increasing the diversity and robustness of the search process. If the second preset condition is the tournament selection method, several groups of parameter configuration schemes are randomly selected for comparison, and the scheme with the highest fitness value is selected to enter the next round. This process is repeated until the required number of schemes are selected. This method ensures that high fitness schemes are given priority through multiple rounds of comparison. If the second preset condition is the random sampling method, a certain number of schemes are randomly selected from all adjusted parameter configuration schemes. This method increases the diversity of the search process.

[0224] By combining different preset fitness conditions, the optical component quality prediction model can more comprehensively select the first parameter configuration scheme, thereby improving the efficiency and effect of the swarm intelligence optimization algorithm, and ultimately improving the quality of the optical component processing process. At the same time, the diversified selection strategy not only ensures the priority of high fitness parameter configuration schemes, but also introduces randomness and diversity to prevent the swarm intelligence optimization algorithm from falling into the local optimal solution, ensuring that the parameter configuration scheme finally selected has better performance in practical applications.

[0225] The optical element production data processing method provided in the embodiment of the present application can ensure the effectiveness and stability of the swarm intelligence optimization algorithm by introducing the adaptive update number and adaptive update radius. At the same time, by introducing the adaptive mutation probability, the diversity of parameter configuration schemes can be increased, and the exploration ability of the algorithm in the search process can be improved. In addition, by adjusting the parameters in the parameter configuration scheme based on the adaptive penalty parameter factor and the adaptive kernel parameter factor, the quality of the solution can be further optimized, ensuring better adaptation to various constraints in a complex optimization environment, thereby improving the overall performance and reliability of the optical element.

[0226] Next, the specific implementation process of step S203 will be discussed in detail.

[0227] Embodiment 5

[0228] Figure 6 A flowchart of Embodiment 5 of a method for processing optical element production data provided by the present application. Based on any of the above embodiments, whether quality problems will occur when processing the optical element to be processed in the optical element processing area is determined based on the predicted thickness data. The process of the method for processing optical element production data may include:

[0229] S601: If the predicted thickness data is not within the preset thickness data range, it is determined that quality problems may occur when the optical element to be processed is processed in the optical element processing area.

[0230] In this step, after the predicted thickness data is obtained through the optical element quality prediction model, the predicted thickness data is compared with a preset thickness data range to determine whether quality problems will occur when the optical element to be processed is processed in the optical element processing area.

[0231] Specifically, if the predicted thickness data falls within the preset thickness data range, it can be determined that there will be no quality problems when the optical element to be processed is processed in the optical element processing area. At this time, the optical element to be processed can be directly processed in the optical element processing area.

[0232] If the predicted thickness data is not within the preset thickness data range, it can be determined that the optical element to be processed may have quality problems when it is processed in the optical element processing area. It is worth noting that after determining that a quality problem occurs, a series of corresponding measures need to be implemented. The specific implementation method can refer to step S701 in the subsequent embodiment 6, which will not be repeated here.

[0233] The preset thickness data range is determined by analyzing a large amount of historical processing data and quality inspection results. The preset thickness data range usually includes the thickness value interval that can ensure the quality of the optical element under normal processing conditions. In this way, the processing quality of the optical element can be pre-judged before actual processing, avoiding processing failure or quality problems caused by unqualified thickness.

[0234] The optical element production data processing method provided in the embodiment of the present application determines whether the optical element to be processed will have quality problems when it is processed in the optical element processing area by comparing the relationship between the predicted thickness data and the preset thickness data range. The method can determine the processing status of the optical element to be processed in the processing area in advance, and take preventive and response measures in time when necessary, so as to ensure the quality of the optical element.

[0235] Embodiment 6

[0236] Figure 7This is a flow chart of Embodiment 6 of a method for processing optical element production data provided by the present application. Based on any of the above embodiments, the process of the method for processing optical element production data also includes:

[0237] S701: If it is determined that quality problems will occur when processing the optical element to be processed in the optical element processing area, the process parameter data of the optical element processing area is automatically adjusted, and quality warning information is pushed to the terminal device of the manager, wherein the quality warning information is used to indicate that quality problems will occur when processing the optical element to be processed in the optical element processing area.

[0238] In this step, if it is determined according to step S601 that the optical element to be processed will have quality problems when being processed in the optical element processing area, the process parameter data of the optical element processing area is automatically adjusted, and quality warning information is pushed to the terminal device of the manager. The quality warning information is used to indicate that the optical element to be processed will have quality problems when being processed in the optical element processing area.

[0239] Specifically, the process parameter data is adjusted based on the difference between the predicted thickness data and the preset thickness data range.

[0240] In a possible implementation, the adjustment of the process parameter data includes but is not limited to adjusting any one or any combination of the temperature data, humidity data, and pressure data of the optical element processing area, as well as the pressure data and optical environment data to which the optical element to be processed is subjected during processing.

[0241] Specifically, temperature data is one of the important factors affecting the processing quality of optical components. If the predicted thickness data indicates that the thickness of the optical component is uneven, the optical component production data processing system will check whether the temperature of the processing area is stable. If the temperature fluctuates greatly, the optical component production data processing system will adjust the control parameters of the heating or cooling device to ensure that the temperature of the processing area remains within a stable range. This can be achieved by increasing or decreasing the heating power, or adjusting the operating frequency of the cooling system, thereby ensuring that the optical component has consistent physical properties during the processing process.

[0242] Humidity data will affect the processing quality of optical components, especially in high-precision processing. If the humidity is too high, it may cause the material to absorb moisture and expand, affecting the processing accuracy. The optical component production data processing system will monitor the humidity data of the processing area and start the dehumidification equipment when necessary to reduce the humidity to an appropriate range. If the humidity is too low, it may cause static electricity accumulation and affect the processing stability. The optical component production data processing system will increase the operating frequency of the humidity control equipment to ensure that the humidity is maintained at a moderate level.

[0243] Pressure data refers to the value that reflects the pressure exerted by the gas or liquid on the unit area in the processing area during the processing of optical components. If the predicted thickness data shows that the thickness of the optical component does not meet expectations, the optical component production data processing system will check the pressure data of the processing area. If the pressure is unstable, the optical component production data processing system will adjust the working parameters of the vacuum pump or pressurizing equipment to ensure that the pressure in the processing area remains within the set range. This can be achieved by adjusting the vacuum pump's exhaust rate or the output pressure of the pressurizing equipment to ensure the stability of the optical component during the processing.

[0244] Pressure data refers to the direct pressure to which the optical component to be processed is subjected during processing. If the predicted thickness data shows that the thickness deviation of the optical component is large, the optical component production data processing system will adjust the pressure applied by the processing equipment. For example, if the thickness is too small, the optical component production data processing system will increase the applied pressure. If the thickness is too large, the optical component production data processing system will reduce the applied pressure. This can be achieved by adjusting the pressure setting value of the hydraulic system or the downward pressure of the mechanical pressure head to ensure that the optical component is subjected to uniform force during processing.

[0245] Optical environment data includes the light intensity and spectral characteristics in the optical component processing area. If the optical environment is not good, the surface quality and optical performance of the optical component may be affected. The optical component production data processing system monitors the optical environment data and adjusts the brightness and spectral distribution of the light source when necessary, or adds filtering equipment to ensure that the optical environment meets the processing requirements. This can be achieved by adjusting the power output of the light source or changing the light source type to ensure that the surface quality and optical performance of the optical component meet the expected standards.

[0246] Through comprehensive adjustment of process parameters such as temperature, humidity, pressure, pressure and optical environment, the optical component production data processing system can ensure that the optical components are always in the optimal state during the processing, thereby improving processing quality and production efficiency.

[0247] It is worth noting that while the optical component production data processing system automatically adjusts the process parameter data of the optical component processing area, it also pushes quality warning information to the terminal device of the manager. Among them, the quality warning information includes but is not limited to any one or any combination of detailed information of the optical component to be processed, predicted thickness data, preset thickness data range, possible types of quality problems and process parameter adjustment suggestions. Specifically, the detailed information of the optical component to be processed includes but is not limited to any one or any combination of number, specification, material, etc. The predicted thickness data is the thickness of the optical component predicted based on the optical component quality prediction model. The preset thickness data range is the thickness range of the optical component that meets the quality requirements. The types of quality problems that may occur include but are not limited to insufficient thickness, excessive thickness, substandard surface roughness, etc. The process parameter adjustment suggestions include but are not limited to specific adjustment measures and expected effects.

[0248] The optical element production data processing method provided in the embodiment of the present application automatically adjusts the process parameter data of the optical element processing area when it is determined that the optical element to be processed will have quality problems in the optical element processing area, and pushes quality warning information to the terminal device of the manager. This method can not only prevent quality problems before they occur, but also respond quickly when quality problems occur, ensuring the final quality of the optical element, thereby improving production efficiency and the quality of the optical element, and reducing production costs and scrap rates.

[0249] In a preferred embodiment, the swarm intelligence optimization algorithm is an enhanced fireworks algorithm to improve the efficiency and accuracy of the search process. At the same time, the machine learning model is a least squares support vector machine model to achieve accurate prediction of the thickness of optical elements. By combining these two methods, the parameter configuration can be optimized more effectively, and the performance and quality of optical elements can be further improved. Among them, the enhanced fireworks algorithm has higher search efficiency and accuracy than other swarm intelligence optimization algorithms, such as particle swarm optimization algorithms or genetic algorithms. The enhanced fireworks algorithm generates diversified solutions in the search space by simulating the behavior of fireworks explosions to avoid falling into local optimality and enhance global search capabilities. When dealing with the optimization of machine learning model parameters, this method can explore possible solutions more comprehensively, so as to find a better parameter configuration, thereby improving the performance and quality of optical elements. Compared with other machine learning models, such as traditional support vector machine models or linear regression models, the least squares support vector machine model provides higher prediction accuracy and stronger generalization ability.

[0250] Specifically, after obtaining the historical thickness data of the optical element to be processed and the historical process parameter data of the optical element processing area, the hyperparameters in the enhanced fireworks algorithm are first initialized. Among them, the hyperparameters in the enhanced fireworks algorithm include but are not limited to I groups of first parameter configuration schemes (corresponding to I groups of population individuals), fireworks explosion intensity (corresponding to the number of adaptive updates), fireworks explosion amplitude (corresponding to the adaptive update radius), upper and lower limits of penalty parameters (corresponding to adaptive penalty parameter factors), upper and lower limits of kernel parameters (corresponding to adaptive kernel parameter factors) and preset iteration termination conditions. Then, I least squares support vector machine models (corresponding to prediction models) with the same structure as the first parameter configuration scheme group are constructed to obtain I first initial optical element quality prediction models. Then the remaining execution steps can refer to steps d to h in the above-mentioned embodiment 3, and will not be repeated here.

[0251] It is worth noting that in the process of improving the quality of optical components, the combination of the enhanced fireworks algorithm and the least squares support vector machine model has significant advantages. This method first obtains the historical thickness data and related process parameter data of the optical components to be processed to provide a solid foundation for subsequent optimization. In this process, the reasonable setting of the hyperparameters of the enhanced fireworks algorithm can effectively improve the efficiency and accuracy of the search process. By constructing multiple least squares support vector machine models with the same number of groups as the first parameter configuration scheme, accurate prediction of the thickness of optical components can be achieved. This feature enables the model to quickly and efficiently fit data when facing complex optical processing processes, so as to better grasp the processing characteristics of optical components. In addition, the superior performance of the least squares support vector machine model in processing small sample data further enhances the ability to optimize the quality of optical components.

[0252] Combining these two methods, the enhanced fireworks algorithm can not only explore a wider range of parameter spaces, but also improve search efficiency through adaptive update strategies. The least squares support vector machine model provides high-precision prediction support for each set of parameter configurations. Ultimately, this optimization combination can achieve a comprehensive improvement in the quality of optical components, making the processing process more stable and reliable, thereby improving production efficiency and reducing costs.

[0253] Device Example 1

[0254] Figure 8 This is a schematic diagram of the structure of a first embodiment of an optical element production data processing device provided by the present application. Figure 8 As shown, the optical element production data processing device 800 includes:

[0255] An acquisition module 801 is used to acquire thickness data of the optical element to be processed and process parameter data of the optical element processing area based on the data uploaded by the sensor device, wherein the process parameter data includes temperature data, humidity data, pressure data of the optical element processing area, and pressure data and optical environment data to which the optical element to be processed is subjected during processing;

[0256] The processing module 802 is used to input the thickness data and process parameter data of the optical element to be processed into a preset optical element quality prediction model to obtain the predicted thickness data of the optical element to be processed when the processing in the optical element processing area is completed. The optical element quality prediction model is obtained by optimizing and training the machine learning model based on the swarm intelligence optimization algorithm in advance, and is used to predict the thickness data of the optical element to be processed when the processing in the optical element processing area is completed.

[0257] The determination module 803 is used to determine whether quality problems will occur when the optical element to be processed is processed in the optical element processing area according to the predicted thickness data.

[0258] Optionally, the processing module 802 is further configured to:

[0259] According to the historical thickness data of the optical components to be processed and the historical process parameter data of the optical component processing area, the swarm intelligence optimization algorithm is used to optimize the parameters of the initial optical component quality prediction model to obtain the optimal parameter configuration solution.

[0260] Optionally, the acquisition module 801 is further used for:

[0261] According to the optimal parameter configuration scheme, the parameters in the initial optical component quality prediction model are configured to obtain the optical component quality prediction model.

[0262] Optionally, the acquisition module 801 is further used for:

[0263] Step a, obtaining a sample data set based on historical thickness data of the optical element to be processed and historical process parameter data of the optical element processing area, the sample data set including multiple groups of data samples, each group of data samples including first historical thickness data of the optical element to be processed before processing, second historical thickness data of the optical element to be processed when processing is completed in the optical element processing area, and historical process parameter data, the historical process parameter data including historical temperature data, historical humidity data, historical pressure data of the optical element processing area, and historical pressure data and historical optical environment data of the optical element to be processed during processing.

[0264] Optionally, the processing module 802 is further configured to:

[0265] Step b, initializing I groups of population individuals, adaptive update quantity, adaptive update radius, adaptive mutation probability, adaptive penalty parameter factor, adaptive kernel parameter factor and preset iteration termination condition, wherein I is a positive integer, each group of population individuals represents a group of first parameter configuration schemes, and each group of first parameter configuration schemes includes a penalty parameter and a kernel parameter;

[0266] Step c: constructing a prediction model with exactly the same structure based on the machine learning model to obtain a first initial optical element quality prediction model;

[0267] Step d: for each first initial optical element quality prediction model, select a set of first parameter configuration schemes, configure the initial parameters of the first initial optical element quality prediction model, and obtain I second initial optical element quality prediction models, wherein the initial parameters of the first initial optical element quality prediction model include an initial penalty parameter and an initial kernel parameter;

[0268] Step e, selecting I groups of sample data from the sample data set, and inputting the first historical thickness data and the historical process parameter data in the I groups of sample data into I second initial optical element quality prediction models for prediction, to obtain I predicted thickness data;

[0269] Step f, calculating the fitness value of each group of first parameter configuration schemes according to I predicted thickness data, the second historical thickness data in I groups of sample data, and the fitness function of the swarm intelligence optimization algorithm;

[0270] Step g, obtaining the first parameter configuration scheme corresponding to the highest fitness value, and recording it as the target parameter configuration scheme;

[0271] Step h: If the preset iteration termination condition is not met, the first parameter configuration scheme is updated according to the fitness value, adaptive update quantity, adaptive update radius, adaptive mutation probability, adaptive penalty parameter factor and adaptive kernel parameter factor of each group of first parameter configuration schemes to obtain I groups of new first parameter configuration schemes. Steps d to h are repeated until the preset iteration termination condition is met, and the target parameter configuration is determined as the optimal parameter configuration scheme.

[0272] Optionally, the processing module 802 is further configured to:

[0273] Calculate the update quantity of each group of first parameter configuration schemes according to the fitness value and the adaptive update quantity of each group of first parameter configuration schemes;

[0274] Calculate the update radius of each group of first parameter configuration schemes according to the fitness value and the adaptive update radius of each group of first parameter configuration schemes;

[0275] According to the update quantity and the update radius, each group of first parameter configuration schemes is updated to obtain a second parameter configuration scheme;

[0276] Randomly selecting a preset number of parameter configuration schemes from the second parameter configuration scheme as a third parameter configuration scheme, and performing a mutation operation on the third parameter configuration scheme according to an adaptive mutation probability to obtain a fourth parameter configuration scheme;

[0277] According to the adaptive penalty parameter factor, the penalty parameter in the second parameter configuration scheme and the penalty parameter in the fourth parameter configuration scheme are adjusted, and according to the adaptive core parameter factor, the core parameter in the second parameter configuration scheme and the core parameter in the fourth parameter configuration scheme are adjusted to obtain the adjusted second parameter configuration scheme and the fourth parameter configuration scheme;

[0278] According to a preset selection strategy, a group I of new first parameter configuration schemes is selected from the adjusted second parameter configuration schemes and the fourth parameter configuration schemes.

[0279] Optionally, the determination module 803 is further configured to:

[0280] If the predicted thickness data is not within the preset thickness data range, it is determined that quality problems will occur when the optical element to be processed is processed in the optical element processing area.

[0281] Optionally, the processing module 802 is further configured to:

[0282] If it is determined that quality problems will occur when processing the optical element to be processed in the optical element processing area, the process parameter data of the optical element processing area is automatically adjusted, and quality warning information is pushed to the terminal device of the manager, wherein the quality warning information is used to indicate that quality problems will occur when processing the optical element to be processed in the optical element processing area.

[0283] Optionally, the adaptive update quantity is calculated using the following formula, including:

[0284]

[0285] in, Indicates The iteration The number of adaptive updates corresponding to the first parameter configuration scheme of the group, Indicates the preset maximum number of updates. Indicates the preset minimum number of updates. Indicates the preset maximum number of iterations, Indicates the current iteration number, Indicates the tth iteration The fitness value corresponding to the first parameter configuration scheme of the group, represents the maximum fitness value at the tth iteration, Indicates a preset constant;

[0286] Accordingly, the adaptive update radius is calculated using the following formula, including:

[0287]

[0288] in, Indicates The iteration The adaptive update radius corresponding to the first parameter configuration scheme of the group, Indicates the preset maximum update radius. Indicates the preset minimum update radius, Represents the minimum fitness value at the tth iteration.

[0289] The optical element production data processing device provided in this embodiment can be used to execute the optical element production data processing method in any of the aforementioned method embodiments, and its implementation principle and technical effects are similar and will not be repeated here.

[0290] Fig. 9 This is a schematic diagram of the structure of an electronic device provided in this application. Fig. 9 As shown, the electronic device 102 may specifically include a receiver 901, a processor 902, and a memory 903. The receiver 901 is used to implement data transmission between the electronic device and the sensor device, and the memory 903 stores computer-executable instructions; the processor 902 executes the computer-executable instructions stored in the memory 903 to implement the optical element production data processing method in the above embodiment.

[0291] Fig.10 This is a schematic diagram of the structure of an optical element production data processing system provided in this application. Fig.10 As shown, the optical element production data processing system 1000 includes Fig. 9 The electronic device 102 and the sensor device 1001 connected to the electronic device 102 are shown;

[0292] The sensor device 1001 is used to collect thickness data of the optical element to be processed and process parameter data of the optical element processing area in real time and upload the collected data to the electronic device 102;

[0293] The sensor device 1001 includes: a spectral confocal sensor 10011, a temperature sensor 10012, a humidity sensor 10013, a pressure sensor 10014, a pressure sensor 10015, and a fiber wave sensor 10016;

[0294] The spectral confocal sensor 10011 is used to collect thickness data of the optical element to be processed;

[0295] The temperature sensor 10012 is used to collect temperature data of the optical element processing area;

[0296] The humidity sensor 10013 is used to collect humidity data in the optical element processing area;

[0297] The pressure sensor 10014 is used to collect pressure data of the optical element processing area;

[0298] The pressure sensor 10015 is used to collect the pressure data of the optical element to be processed during processing;

[0299] The fiber optic wave sensor 10016 is used to collect optical environment data of the optical element to be processed during processing.

[0300] This embodiment provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the optical element production data processing method of any one of the above embodiments.

[0301] This embodiment further provides a computer program product, including a computer program, which, when executed by a processor, implements the optical element production data processing method provided in any one of the above embodiments.

[0302] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary techniques in the art that are not disclosed in the present application. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present application are indicated by the following claims.

[0303] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A method for processing optical element production data, characterized in that: include: Based on the data uploaded by the sensor device, the thickness data of the optical element to be processed and the process parameter data of the optical element processing area are obtained, wherein the process parameter data include the temperature data, humidity data, pressure data of the optical element processing area, and the pressure data and optical environment data to which the optical element to be processed is subjected during processing; Inputting the thickness data of the optical element to be processed and the process parameter data into a preset optical element quality prediction model to obtain predicted thickness data of the optical element to be processed when processing in the optical element processing area is completed, wherein the optical element quality prediction model is obtained by optimizing and training a machine learning model based on a swarm intelligence optimization algorithm in advance, and is used to predict the thickness data of the optical element to be processed when processing in the optical element processing area is completed; Determining, based on the predicted thickness data, whether quality problems will occur when the optical element to be processed is processed in the optical element processing area; The process of obtaining the optical element quality prediction model includes: According to the historical thickness data of the optical element to be processed and the historical process parameter data of the optical element processing area, the swarm intelligence optimization algorithm is used to optimize the parameters of the initial optical element quality prediction model to obtain the optimal parameter configuration scheme; According to the optimal parameter configuration scheme, the parameters in the initial optical element quality prediction model are configured to obtain the optical element quality prediction model; The method uses the swarm intelligence optimization algorithm to optimize the parameters of the initial optical element quality prediction model according to the historical thickness data of the optical element to be processed and the historical process parameter data of the optical element processing area to obtain the optimal parameter configuration scheme, including: Step a, acquiring a sample data set based on the historical thickness data of the optical element to be processed and the historical process parameter data of the optical element processing area, wherein the sample data set includes multiple groups of data samples, each group of data samples includes first historical thickness data of the optical element to be processed before processing, second historical thickness data of the optical element to be processed when processing is completed in the optical element processing area, and the historical process parameter data, wherein the historical process parameter data includes historical temperature data, historical humidity data, historical pressure data of the optical element processing area, and historical pressure data and historical optical environment data of the optical element to be processed during processing; Step b, initializing I groups of population individuals, adaptive update quantity, adaptive update radius, adaptive mutation probability, adaptive penalty parameter factor, adaptive kernel parameter factor and preset iteration termination condition, wherein I is a positive integer, each group of population individuals represents a group of first parameter configuration schemes, and each group of first parameter configuration schemes includes a penalty parameter and a kernel parameter; Step c: constructing a prediction model with exactly the same structure based on the machine learning model to obtain a first initial optical element quality prediction model; Step d: for each first initial optical element quality prediction model, select a set of first parameter configuration schemes, configure the initial parameters of the first initial optical element quality prediction model, and obtain I second initial optical element quality prediction model, wherein the initial parameters of the first initial optical element quality prediction model include an initial penalty parameter and an initial kernel parameter; Step e, selecting I groups of sample data from the sample data set, and inputting the first historical thickness data and the historical process parameter data in the I groups of sample data into I second initial optical element quality prediction models for prediction, to obtain I predicted thickness data; Step f, calculating the fitness value of each group of first parameter configuration schemes according to the I predicted thickness data, the second historical thickness data in the I group of sample data, and the fitness function of the swarm intelligence optimization algorithm; Step g, obtaining the first parameter configuration scheme corresponding to the highest fitness value, and recording it as the target parameter configuration scheme; Step h: If the preset iteration termination condition is not met, the first parameter configuration scheme is updated according to the fitness value of each group of first parameter configuration schemes, the adaptive update quantity, the adaptive update radius, the adaptive mutation probability, the adaptive penalty parameter factor and the adaptive kernel parameter factor to obtain I groups of new first parameter configuration schemes, and steps d to h are repeated until the preset iteration termination condition is met, and the target parameter configuration is determined as the optimal parameter configuration scheme.

2. The method according to claim 1, characterized in that The updating of the first parameter configuration scheme according to the fitness value of each group of first parameter configuration schemes, the adaptive update quantity, the adaptive update radius, the adaptive mutation probability, the adaptive penalty parameter factor, and the adaptive kernel parameter factor to obtain a group of new first parameter configuration schemes includes: Calculate the update quantity of each group of first parameter configuration schemes according to the fitness value of each group of first parameter configuration schemes and the adaptive update quantity; Calculating the update radius of each group of first parameter configuration schemes according to the fitness value of each group of first parameter configuration schemes and the adaptive update radius; According to the update quantity and the update radius, each group of first parameter configuration schemes is updated to obtain a second parameter configuration scheme; randomly selecting a preset number of parameter configuration schemes from the second parameter configuration schemes as third parameter configuration schemes, and performing a mutation operation on the third parameter configuration scheme according to the adaptive mutation probability to obtain a fourth parameter configuration scheme; According to the adaptive penalty parameter factor, adjusting the penalty parameter in the second parameter configuration scheme and the penalty parameter in the fourth parameter configuration scheme, and according to the adaptive core parameter factor, adjusting the core parameter in the second parameter configuration scheme and the core parameter in the fourth parameter configuration scheme to obtain adjusted second parameter configuration scheme and fourth parameter configuration scheme; According to a preset selection strategy, a group I of new first parameter configuration schemes is selected from the adjusted second parameter configuration schemes and the fourth parameter configuration schemes.

3. The method according to any one of claims 1 or 2, characterized in that: Determining, based on the predicted thickness data, whether quality problems will occur when processing the optical element to be processed in the optical element processing area comprises: If the predicted thickness data is not within the preset thickness data range, it is determined that quality problems will occur when processing the optical element to be processed in the optical element processing area.

4. The method according to claim 3, characterized in that The method further comprises: If it is determined that quality problems will occur when processing the optical element to be processed in the optical element processing area, the process parameter data of the optical element processing area is automatically adjusted, and quality warning information is pushed to the terminal device of the administrator, wherein the quality warning information is used to indicate that quality problems will occur when processing the optical element to be processed in the optical element processing area.

5. The method according to claim 1, characterized in that The adaptive update quantity is calculated using the following formula, including: in, Indicates The first iteration The number of adaptive updates corresponding to the first parameter configuration scheme of the group, Indicates the preset maximum number of updates. Indicates the preset minimum number of updates. Indicates the preset maximum number of iterations, Indicates the current iteration number, Indicates the tth iteration The fitness value corresponding to the first parameter configuration scheme of the group, represents the maximum fitness value at the tth iteration, Indicates a preset constant; Accordingly, the adaptive update radius is calculated using the following formula, including: in, Indicates The first iteration The adaptive update radius corresponding to the first parameter configuration scheme of the group, Indicates the preset maximum update radius. Indicates the preset minimum update radius, Represents the minimum fitness value at the tth iteration.

6. An optical element production data processing device, characterized in that: include: An acquisition module, used to acquire thickness data of the optical element to be processed and process parameter data of the optical element processing area based on the data uploaded by the sensor device, wherein the process parameter data includes temperature data, humidity data, pressure data of the optical element processing area, and pressure data and optical environment data to which the optical element to be processed is subjected during processing; a processing module, for inputting the thickness data of the optical element to be processed and the process parameter data into a preset optical element quality prediction model, to obtain predicted thickness data of the optical element to be processed when processing in the optical element processing area is completed, wherein the optical element quality prediction model is obtained by optimizing and training a machine learning model based on a swarm intelligence optimization algorithm in advance, and is a model for predicting thickness data of the optical element to be processed when processing in the optical element processing area is completed; A determination module, configured to determine, based on the predicted thickness data, whether quality problems will occur when processing the optical element to be processed in the optical element processing area; The processing module is further used for: According to the historical thickness data of the optical element to be processed and the historical process parameter data of the optical element processing area, the swarm intelligence optimization algorithm is used to optimize the parameters of the initial optical element quality prediction model to obtain the optimal parameter configuration scheme; The acquisition module is further used to: configure the parameters in the initial optical element quality prediction model according to the optimal parameter configuration scheme to obtain the optical element quality prediction model; The acquisition module is further used for: step a, acquiring a sample data set based on the historical thickness data of the optical element to be processed and the historical process parameter data of the optical element processing area, the sample data set includes a plurality of groups of data samples, each group of data samples includes the first historical thickness data of the optical element to be processed before processing, the second historical thickness data of the optical element to be processed when processing is completed in the optical element processing area, and the historical process parameter data, the historical process parameter data includes the historical temperature data, historical humidity data, historical pressure data of the optical element processing area, and the historical pressure data and historical optical environment data of the optical element to be processed during processing; The processing module is further used for: step b, initializing I groups of population individuals, adaptively updating the number, adaptively updating the radius, adaptively changing the probability, adaptively punishing the parameter factor, adaptively punishing the kernel parameter factor and presetting the iteration termination condition, wherein I is a positive integer, each group of population individuals represents a group of first parameter configuration schemes, and each group of first parameter configuration schemes includes a punishment parameter and a kernel parameter; Step c: constructing a prediction model with exactly the same structure based on the machine learning model to obtain a first initial optical element quality prediction model; Step d: for each first initial optical element quality prediction model, select a set of first parameter configuration schemes, configure the initial parameters of the first initial optical element quality prediction model, and obtain I second initial optical element quality prediction model, wherein the initial parameters of the first initial optical element quality prediction model include an initial penalty parameter and an initial kernel parameter; Step e, selecting I groups of sample data from the sample data set, and inputting the first historical thickness data and the historical process parameter data in the I groups of sample data into I second initial optical element quality prediction models for prediction, to obtain I predicted thickness data; Step f, calculating the fitness value of each group of first parameter configuration schemes according to the I predicted thickness data, the second historical thickness data in the I group of sample data, and the fitness function of the swarm intelligence optimization algorithm; Step g, obtaining the first parameter configuration scheme corresponding to the highest fitness value, and recording it as the target parameter configuration scheme; Step h: If the preset iteration termination condition is not met, the first parameter configuration scheme is updated according to the fitness value of each group of first parameter configuration schemes, the adaptive update quantity, the adaptive update radius, the adaptive mutation probability, the adaptive penalty parameter factor and the adaptive kernel parameter factor to obtain I groups of new first parameter configuration schemes, and steps d to h are repeated until the preset iteration termination condition is met, and the target parameter configuration is determined as the optimal parameter configuration scheme.

7. An electronic device, characterized in that: include: A receiver, a processor and a memory connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the optical element production data processing method according to any one of claims 1 to 5.

8. An optical element production data processing system, characterized in that: include: The electronic device according to claim 7 and a sensor device connected to the electronic device; The sensor device is used to collect thickness data of the optical element to be processed and process parameter data of the optical element processing area in real time and upload the collected data to the electronic device; The sensor equipment includes: a spectral confocal sensor, a temperature sensor, a humidity sensor, a pressure sensor, a pressure sensor, and a fiber wave sensor; The spectral confocal sensor is used to collect thickness data of the optical element to be processed; The temperature sensor is used to collect temperature data of the optical element processing area; The humidity sensor is used to collect humidity data of the optical element processing area; The pressure sensor is used to collect pressure data of the optical element processing area; The pressure sensor is used to collect pressure data on the optical element to be processed during processing; The fiber optic wave sensor is used to collect optical environment data of the optical element to be processed during processing.

Images