Optical system design method and system based on knowledge base retrieval enhanced large model

Abstract

The application discloses an optical system design method and system based on a knowledge base retrieval enhanced large model, and belongs to the technical field of optical design and artificial intelligence. The method comprises the following steps: constructing an optical system knowledge base containing vectorized and encoded optical lens design data; constructing data pairs based on optical system design requirements and corresponding parameter lists of surfaces of the optical system, and constructing a training data set; training a large language model by using the training data set and adopting a low-rank adaptive technology; combining the trained large language model and the optical system knowledge base through a knowledge base retrieval enhancement module to obtain a knowledge base retrieval enhanced large model; inputting design requirements of an optical system to be designed into the knowledge base retrieval enhanced large model to obtain optical system structure parameters meeting the design requirements of the optical system to be designed, i.e., parameter lists of surfaces of the optical system. The method can effectively improve the optical design efficiency and is suitable for the design of various optical systems.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of optical design automation, and particularly relates to an optical system design method and system based on knowledge base retrieval enhanced large model. BACKGROUND

[0002] Optical systems are the core components in many high-tech fields such as imaging, illumination, and sensing. The traditional optical system design process highly depends on the professional knowledge and long-term accumulated experience of the design engineers, and usually includes steps such as selecting an initial structure according to the requirements, using optical design software (such as Zemax, CODE V) for multiple rounds of parameter optimization and image quality evaluation. This process is not only time-consuming and laborious, but also for complex systems or innovative structures, the designer often has difficulty in quickly finding a suitable initial structure, resulting in low design efficiency and long innovation cycle.

[0003] In recent years, artificial intelligence technologies such as deep learning have shown great potential in many fields. In the field of optical design, there have been attempts to use neural networks to directly infer optical structures or optimize parameters. For example, Paszke et al. proved that deep neural networks (DNNs) can be used to generate initial optical structures; Cote et al. applied recurrent neural networks (RNNs) to the inference of lens design sequences. However, these methods usually face the following challenges:

[0004] Optimization target is single: existing methods mostly focus on optimizing a single or a few optical performance indicators (such as RMS (Root Mean Square) spot size), and it is difficult to take into account the complex multi-objective constraints in actual design (such as distortion, MTF (Modulation Transfer Function), size, cost, manufacturability, etc.).

[0005] Weak structure discovery ability: the model mainly performs interpolation or combination on existing design data, and it is difficult to actively discover new and breakthrough optical structures.

[0006] Insufficient use of prior knowledge: unable to fully combine optical design principles, formulas and empirical rules to guide the model learning and generation process.

[0007] Large language models (LLMs) such as the GPT (Generative Pre-trained Transformer) series, Qwen (commonly known as Qwen) series, and the like, with their powerful text understanding, generation, and zero-shot / few-shot learning capabilities, provide a new way to solve the above problems. Through fine-tuning and combination with external knowledge bases (such as Retrieval Augmented Generation, RAG), LLMs are expected to understand complex optical design requirements and generate initial optical structures that meet the requirements. Currently, the application of LLMs, especially in combination with RAG technology, to optical system design is still in its infancy.

[0008] Therefore, it is of great significance to develop a method and system that can effectively utilize existing optical design knowledge, understand user requirements, and intelligently generate high-quality initial optical system structures, to improve optical design efficiency and promote optical technology innovation. SUMMARY

[0009] To solve the problems in the prior art, the present application provides an optical system design method and system based on a knowledge base retrieval augmented large model.

[0010] The technical solutions of the present application are as follows:

[0011] In a first aspect, the present application discloses an optical system design method based on a knowledge base retrieval augmented large model, comprising:

[0012] 1) Obtain optical lens design data and preprocess it, vectorize and encode the preprocessed optical lens design data, and construct an optical system knowledge base;

[0013] 2) Obtain a parameter list of each surface of the optical system corresponding to the optical system design requirements, construct a design requirement-parameter list data pair, and then construct a training data set based on the design requirement-parameter list data pair;

[0014] 3) Train the large language model using the training data set and a low-rank adaptive method to obtain an optical system design large model; then combine the optical system design large model with the optical system knowledge base through a knowledge base retrieval augmentation module to obtain a knowledge base retrieval augmented large model composed of the knowledge base retrieval augmentation module and the optical system design large model;

[0015] 4) Obtain the design requirements of the optical system to be designed and input them into the knowledge base retrieval augmentation module for information retrieval to obtain a number of optical design information most relevant to the design requirements; then input the design requirements and the optical design information into the optical system design large model to obtain the optical system structure parameters that meet the design requirements of the optical system to be designed, i.e., the parameter list of each surface of the optical system.

[0016] In a second aspect, the present application discloses an optical system design system based on a knowledge base retrieval enhanced large model for realizing the design method, comprising:

[0017] a knowledge base construction module for acquiring optical lens design data and preprocessing, vectorizing and encoding the preprocessed optical lens design data, and constructing an optical system knowledge base;

[0018] a data set construction module for acquiring a parameter list of each surface of the optical system corresponding to the optical system design requirements, constructing a design requirement-parameter list data pair, and constructing a training data set based on the design requirement-parameter list data pair;

[0019] a model construction module for training a large language model using the training data set and adopting a low-rank adaptive method to obtain an optical system design large model, and combining the optical system design large model with the optical system knowledge base through a knowledge base retrieval enhancement module to obtain a knowledge base retrieval enhanced large model composed of the knowledge base retrieval enhancement module and the optical system design large model;

[0020] a design module for acquiring design requirements of an optical system to be designed, inputting the design requirements into the knowledge base retrieval enhancement module for information retrieval to obtain a plurality of optical design information most relevant to the design requirements, and inputting the design requirements and the optical design information into the optical system design large model to obtain the structure parameters of the optical system meeting the design requirements of the optical system to be designed, i.e., the parameter list of each surface of the optical system.

[0021] Compared with the prior art, the present application has the following advantages:

[0022] 1) Overcoming the single optimization target, realizing the understanding of complex multi-objective design requirements: In view of the pain point that the existing method is difficult to consider multiple target constraints such as distortion, MTF, size, cost, etc., the present application constructs an "optical system design requirement" template containing multi-dimensional indicators, format requirements and even optical formulas, and uses the powerful text understanding ability of the large language model to comprehensively understand and process complex multi-objective design tasks described in natural language, rather than only optimizing a single RMS spot size, so that the generated results are closer to actual engineering requirements.

[0023] 2) Enhancing the structure discovery ability and promoting design innovation: In view of the pain point that the existing method mainly interpolates or combines on existing data and is difficult to discover new structures, the present application uses the knowledge base retrieval enhancement (RAG) technology to retrieve multiple related but possibly different design cases from a large knowledge base as references before generation. The large language model performs reasoning and fusion on this basis, rather than simply imitating, which can jump out of local optimization and has the potential to generate breakthrough and non-intuitive new optical structures, thereby enhancing the innovation of the design.

[0024] 3) Deeply utilize prior knowledge to improve the rationality of the generated scheme: To address the pain point of insufficient utilization of optical principles and empirical rules by existing methods, the invention builds a knowledge base containing vast amounts of optical design data, formulas, and rules. Through RAG technology, the most relevant prior knowledge (such as the application scope of specific materials, the design paradigm of classic structures, Abbe invariant, etc.) is dynamically injected into the context of the large model during the generation process, guiding the model to generate an initial structure that is physically reasonable and conforms to engineering practice, significantly improving the usability and quality of the scheme.

[0025] 4) Automation and efficiency improvement, reducing the experience threshold: Combining the above advantages, the invention models the design ideas and knowledge base of experienced optical engineers, automatically generating high-quality initial optical structures through large models, significantly reducing the time for manual brainstorming and trial and error, and greatly improving design efficiency. This allows inexperienced designers to quickly obtain reasonable and even innovative design starting points, reducing dependence on experienced engineers and accelerating the entire product iteration cycle. BRIEF DESCRIPTION OF DRAWINGS

[0026] Figure 1 is a flowchart of the optical system design method based on knowledge base retrieval and enhanced large model of the invention;

[0027] Figure 2 is a structural diagram of the optical system design system based on knowledge base retrieval and enhanced large model of the invention;

[0028] Figure 3 is an optical path diagram of the optical system generated by the optical system design method of the embodiment of the invention;

[0029] Figure 4 is an optical path diagram of the optical system generated by the existing method;

[0030] Figure 5 is an optical path diagram of the optical system generated by the invention under the optical system design requirements of Figure 4 . DETAILED DESCRIPTION

[0031] The invention will be further described and explained with specific embodiments. The embodiments are only exemplary and do not limit the scope of the disclosure. The technical features of each embodiment of the invention can be combined accordingly without conflict.

[0032] The main purpose of the invention is to overcome the shortcomings of the prior art, provide an optical system design method and system based on knowledge base retrieval and enhanced large model, and realize the automatic, intelligent, and efficient generation of optical system structure parameters.

[0033] To achieve the above object, the technical scheme adopted by the present application is as follows:

[0034] The present application provides an optical system design method and system based on knowledge base retrieval enhanced large model, the core of the optical system design method is to understand the user's optical design demand and intelligently generate the initial optical system structure by constructing the optical system knowledge base, fine-tuning the large language model and combining the retrieval enhancement generation technology. Figure 1 As shown in the figure, the method mainly includes the following steps:

[0035] (1) Construction of optical system knowledge base

[0036] This step aims to establish a comprehensive and structured optical system knowledge base, which is in a form that can be efficiently retrieved and utilized by a large language model.

[0037] (1.1) Collection and preprocessing of optical lens design data: Collect and organize a large amount of optical lens design data from various sources, such as extracting from lens library files in ZMX, TXT or XLSX formats containing a large amount of optical lens design data. A typical lens library file can contain about 2 million optical lens design data. Each optical lens design data contains optical system parameters and / or surface parameters, optical system parameters include but are not limited to field of view (FOV), F number (f num ), aperture surface index, effective focal length (EFL), RMS spot size (fit), aperture type, aperture value, field of view number and wavelength, etc.; surface parameters include but are not limited to serial number, type (such as STANDARD), remarks (such as OBJ, STOP), curvature radius (mm), thickness (mm), material, refractive index, Abbe number and half radius (mm) of each optical surface.

[0038] Then, the collected optical lens design data is preprocessed, that is, the collected optical lens design data is cleaned and formatted. Cleaning refers to filtering and deduplicating the optical lens design data, filtering means removing low-quality optical lens design data; low-quality optical lens design data generally refers to optical lens design data with large RMS spot size. Format unification refers to organizing the cleaned optical lens design data into a structured or semi-structured form (such as TXT, HTML, MarkDown or JSON file). For example, all cleaned optical lens design data can be aggregated and stored as a structured JSON file to preserve the hierarchical relationship between parameters, or stored as a plain text (TXT) file separated by a specific delimiter (such as "===").

[0039] (1.2) Vectorization encoding of pre-processed optical lens design data and construction of optical system knowledge base: In order to enable large language models to understand and retrieve information in the knowledge base, it is necessary to convert the pre-processed optical lens design data into vector representation. The present application uses advanced text embedding models such as BGE-M3 to vectorize the pre-processed optical design data and generate high-dimensional feature vectors for subsequent semantic retrieval. These high-dimensional feature vectors are then stored in a dedicated optical system knowledge base to support subsequent fast retrieval based on semantic similarity. The optical system knowledge base may, for example, contain more than 2 million high-dimensional feature vectors obtained by vectorizing the optical lens design data.

[0040] (2) Construction of training data set for fine-tuning of large language model

[0041] This step aims to create a special data set that can train a large language model to understand optical design requirements and generate corresponding structural parameters.

[0042] (2.1) Generation of design requirement-parameter list data pairs: Using existing optical design files (such as Zemax file library), detailed optical parameters are extracted in batches through automated scripts (such as ZPL macros or Python scripts combined with Zemax API). At the same time, simulate the design requests that a real user may propose, and construct the optical system design requirements. Each optical system design requirement contains optical system parameters (i.e. target optical system parameters such as FOV=10.0, f num=2.0, fit = 0.03370, aperture is surface.1, etc.), and the format requirements for the optical system structure parameters of the large language model output (such as please give me detailed parameters of every surfaces in form of markdown), such as the format requirements for the optical system structure parameters of the large language model output, for example, output the optical system structure parameters in the form of Markdown table, HTML table, XLSX table, or JSON format. The design requirements section can also supplement the definition of optical system parameters (i.e. the definition of each component within the optical system parameters), the imaging quality requirements of the optical system at different field angles, and the imaging quality requirements of the optical system at different wavelengths at different field angles. In order to enhance the understanding ability and generation quality of the large language model, the optical design related basic formulas (such as numerical aperture, Snell's law, Abbe invariant, Lagrange invariant, etc.) and guiding prompts and constraints for the model generation process (such as ensuring the readability of the output table, preferentially selecting common optical glass materials, ensuring that light can pass through all generated surfaces, etc.) can be embedded in the optical system design requirements to guide the trained large language model to generate more reasonable designs.

[0043] The parameter list section is the detailed parameter list of each surface of the optical system corresponding to the optical system design requirements, which meets the requirements. It is equivalent to the optical system structure parameters. It is usually presented in structured forms such as Markdown table, HTML table, and JSON text.

[0044] (2.2) Format of data pairs: Organize the extracted optical system structure parameters (as parameter lists) according to the format specified in the optical system design requirements (such as Markdown table). Each pair of optical system design requirements and its corresponding parameter list constitutes a training sample. Through this method, hundreds of such data pairs are generated, and they are organized into a format suitable for fine-tuning of large language models (such as JSON files), and then a training data set for fine-tuning of large language models is formed.

[0045] (3) Efficient fine-tuning of large language model parameters

[0046] This step aims to adapt the general pre-trained large language model to the field of optical design, so that it can generate optical system structure parameters according to specific optical system design requirements.

[0047] (3.1) Selection of basic large language model: Select a pre-trained large language model with strong text understanding and generation capabilities as the basis, such as Qwen2.5-14B-Instruct model.

[0048] (3.2) Implementation of parameter-efficient fine-tuning: Adopt parameter-efficient fine-tuning techniques such as Low Rank Adaptation (LoRA) to fine-tune the base large language model using the training dataset constructed in step (2) based on the design requirement-parameter list data. The fine-tuning process is usually performed on cloud service platforms (such as SiliconCloud) with corresponding computing resources or local computing resources (such as servers with one or more GPUs).

[0049] During the fine-tuning process, a series of hyperparameters need to be carefully adjusted, including learning rate, epochs, batch size, specific parameters of LoRA technology (such as LoRA rank, LoRA Alpha value, Dropout rate), and the maximum number of tokens (max tokens) of the input sequence. By optimizing these parameters, the fine-tuned large language model can accurately understand the optical system design requirements and generate optical system structure parameters that meet the format and content requirements. After fine-tuning, an optimized large model for the optical design field is obtained, i.e., an optical system design large model is obtained. For example, a model named ft:LoRA / Qwen / Qwen2.5-14B-Instruct can be obtained.

[0050] During the parameter-efficient fine-tuning of the large language model, the cross-entropy loss function can be used as the loss function.

[0051] (4) Integration and application of knowledge base retrieval augmented generation (RAG) module:

[0052] This step aims to combine the fine-tuned large language model with the constructed optical system knowledge base to improve the relevance and accuracy of the model's generated results.

[0053] (4.1) Configuration of RAG framework: Use mature retrieval augmented generation frameworks such as RAGFlow or similar technologies such as DIFY. Connect the optical system knowledge base in step (1.2) to the retrieval module of the RAGFlow framework. At the same time, configure the service interface (such as endpoint address, API key, and model name compatible with OpenAI API specifications) of the optical system design large model obtained in step (3.2) to the knowledge base retrieval augmented generation module of the RAGFlow framework, i.e., obtain a knowledge base retrieval augmented large model composed of a knowledge base retrieval augmented module and an optical system design large model. When the knowledge base retrieval augmented large model receives input, the knowledge base retrieval augmented generation module can first retrieve one or more of the most relevant optical system design instances, surface parameter ranges, or optical system design rules from the optical system knowledge base.

[0054] (4.2) Receiving and processing of the optical system design requirements to be designed: The knowledge base retrieval augmented large model receives the optical system design requirements to be designed. The format and content of the optical system design requirements to be designed should be consistent with the format of the simulated real optical system design requirements in step (2.1), so as to ensure that the optical system design large model of the knowledge base retrieval augmented large model and the knowledge base retrieval augmented generation module can accurately understand.

[0055] (4.3) Knowledge retrieval and context augmentation based on RAG: When the knowledge base retrieval augmented generation module receives the optical system design requirements to be designed, its retriever component first vectorizes the optical system design requirements to be designed, and then performs a semantic similarity search (for example, calculates the cosine similarity) in the optical system knowledge base constructed in step (1.2) to find out a number of optical system design cases, surface parameter ranges, optical system design rules or related formula information pieces that are most relevant to the current optical system design requirements to be designed, that is, to find out the optical design information. These retrieved information pieces will serve as the context of the optical system design requirements to be designed.

[0056] (4.4) Structure generation based on augmented context: The knowledge base retrieval augmented generation module inputs the generated optical design information into the optical system design large model of the knowledge base retrieval augmented large model, and at the same time, the optical system design large model of the knowledge base retrieval augmented large model also receives the optical system design requirements to be designed. The optical system design requirements to be designed and the optical design information will form a more content-rich and more specific composite prompt (augmented prompt), and the large model will perform reasoning and content generation based on this augmented context, and finally output the optical system structure parameters that meet the optical system design requirements to be designed, and present them in the format required in the optical system design requirements to be designed (such as Markdown table, HTML table, JSON structured text, etc.). The optical system structure parameters are the parameter list of each surface of the optical system, which includes the surface shape, thickness, material, or refractive index of each surface of the optical system.

[0057] Through the above steps, the present application realizes a method that can combine large-scale prior knowledge and real-time user requirements to intelligently generate optical system structures.

[0058] Implementation Example

[0059] In order to verify the effectiveness of the method described in the present application, experiments were conducted on a server (for example, a cloud server instance with NVIDIA GPU) configured with appropriate computing resources.

[0060] 1. Experimental environment and parameter settings:

[0061] Optical system knowledge base: An XLSX file set containing about 2 million pieces of optical lens design data was used. After preprocessing, the BGE-M3 model was used for vectorization coding and stored in the optical system knowledge base.

[0062] Optical system design large model: Qwen2.5-14B-Instruct was selected as the basic large language model for fine-tuning.

[0063] Fine-tuned training data set: By parsing about 1000 Zemax design files, about 1000 pairs of design requirement-parameter list data pairs were automatically generated and formatted into JSONL files.

[0064] Fine-tuning parameters (on local computing resources): The learning rate is set to 0.0001, the number of training iterations (epochs) is 3, the batch size is 8, the LoRA rank is 8, the LoRA Alpha value is 32, the Dropout rate is 0.05, and the maximum Token number is 32768. After successful fine-tuning, the model version ft:LoRA / Qwen / Qwen2.5-14B-Instruct is obtained.

[0065] Knowledge base retrieval enhanced large model: Built using the RAGFlow framework, the above optical system knowledge base and fine-tuned optical system design large model are configured, that is, the optical system design large model is combined with the optical system knowledge base through retrieval enhancement generation technology to obtain a knowledge base retrieval enhanced large model composed of a knowledge base retrieval enhanced module and an optical system design large model.

[0066] 2. Generation of optical system structure parameter examples: The method of the present application can generate optical system structure parameters containing multiple lenses according to different optical system design requirements to be designed (such as FOV=10, f num =2, diaphragm located in front of the first lens). The generated optical system structure parameters can be directly imported into optical design software such as Zemax for further image quality evaluation and optimization. As shown in Figure 3 , it is the optical path diagram of an optical system obtained by the optical system design method of the present application.

[0067] The specification requirements given to the model are: half field of view (HFOV) = 40 degrees, F number (F / #) = 4, effective focal length (EFL) = 50 mm, and the working wavelength is three colors (656 nm, 588 nm, 486 nm). At the same time, the model is required to include the stop and image surface in the design, and the medium in front of the image surface should be air, and the image surface is located at the best imaging point.

[0068] As a comparison, Figure 4 The optical system generated by the existing deep learning method under the same specification requirements is shown. The system is a relatively simple double-cemented lens group. From its optical path diagram, it can be seen that light rays of different colors disperse seriously near the image surface and fail to converge into a clear focal point, which indicates that the system has significant chromatic aberration and other aberrations, and the imaging quality is poor, which is difficult to meet the design requirements.

[0069] Figure 5 The optical system generated by the method of the present application under the same specification requirements is shown. Compared with Figure 4 the system generated by the present application is a more complex multi-piece lens group. From its optical path diagram, it can be clearly seen that light rays of different colors and different fields of view can converge into a very small diffraction spot after passing through the system, and the focal point is clear and sharp. This indicates that the system generated by the present application effectively corrects chromatic aberration and other aberrations, and its optical performance is better than that of the system generated by the existing method. This comparison fully proves that the present application can generate more complex, more optimized, and more in line with high-performance design requirements initial structure of optical system by combining large-scale knowledge base and reasoning ability of large language model.

[0070] As Figure 2 shown in the embodiments of the present application, the present application also provides an optical system design system based on knowledge base retrieval enhanced large model for implementing the design method, which includes a knowledge base construction module, a data set construction module, a model construction module, and a design module.

[0071] The knowledge base construction module is used for acquiring optical lens design data and pre-processing, vectorizing and coding the pre-processed optical lens design data, and constructing an optical system knowledge base. The data set construction module is used for acquiring a parameter list of each surface of the optical system corresponding to the optical system design requirement, constructing a design requirement-parameter list data pair, and constructing a training data set based on the design requirement-parameter list data pair. The model construction module is used for training a large language model by using the training data set and adopting a low-rank adaptive method, obtaining an optical system design large model, combining the optical system design large model with the optical system knowledge base through a knowledge base retrieval enhancement module, and obtaining a knowledge base retrieval enhancement large model composed of the knowledge base retrieval enhancement module and the optical system design large model. The design module is used for acquiring a design requirement of an optical system to be designed, inputting the design requirement into the knowledge base retrieval enhancement module for information retrieval, obtaining a plurality of optical design information most relevant to the design requirement, inputting the design requirement and the optical design information into the optical system design large model, and obtaining an optical system structure parameter meeting the design requirement of the optical system to be designed, i.e., obtaining a parameter list of each surface of the optical system.

[0072] The application solves the problem of excessive dependence on artificial experience in traditional optical design by the intelligent generation mechanism guided by knowledge, provides an automatic initial scheme generation capability for optical system design, effectively improves the optical design efficiency, and is suitable for innovative design of various imaging optical systems.

[0073] The above-described embodiments only express several embodiments of the application, and the description is more specific and detailed, but it cannot be understood as a limitation on the scope of the patent of the application. For ordinary skilled persons in the art, without departing from the concept of the application, a number of modifications and improvements can be made, which are all within the protection scope of the application.

Claims

1. An optical system design method based on knowledge base retrieval and enhanced large model, characterized in that: include: 1) Obtain and preprocess optical lens design data, vectorize and encode the preprocessed optical lens design data, and build an optical system knowledge base; 2) Obtain a parameter list of each surface of the optical system corresponding to the optical system design requirements, construct a design requirement-parameter list data pair, and then construct a training dataset based on the design requirement-parameter list data pair; 3) Using the training dataset and a low-rank adaptive method to train the large language model, a large model for optical system design is obtained; Then, the optical system design large model is combined with the optical system knowledge base through the knowledge base retrieval enhancement module to obtain a knowledge base retrieval enhancement large model consisting of the knowledge base retrieval enhancement module and the optical system design large model; 4) Obtain the design requirements of the optical system to be designed and input them into the knowledge base retrieval enhancement module for information retrieval to obtain the optical design information most relevant to the design requirements. Then, input the design requirements and optical design information together into the optical system design model to obtain the optical system structural parameters that meet the design requirements of the optical system to be designed, that is, obtain a parameter list of each surface of the optical system.

2. The optical system design method based on knowledge base retrieval and enhanced large model according to claim 1 is characterized in that: In step 1), optical lens design data is obtained from an existing lens library file, wherein the optical lens design data includes optical system parameters and / or surface parameters, wherein the optical system parameters include field of view angle, F number, aperture surface index, effective focal length, RMS spot size, aperture type, aperture value, field of view number and wavelength, and the surface parameters include the serial number, type, remarks, curvature radius, thickness, material, refractive index, Abbe number and semi-aperture of each optical surface.

3. The optical system design method based on knowledge base retrieval and enhanced large model according to claim 1, characterized in that: The pretreatment in step 1) includes: First, the optical lens design data is screened and deduplicated, and then the format of the screened and deduplicated optical lens design data is unified to complete the preprocessing; The vectorized encoding is: using the BGE-M3 model to vectorize the pre-processed optical lens design data.

4. The optical system design method based on knowledge base retrieval and enhanced large model according to claim 1, characterized in that: In step 2), the optical system design requirements include optical system parameters, definitions of optical system parameters, requirements for optical system imaging quality at various field of view angles, requirements for optical system imaging quality at different wavelengths at various field of view angles, format requirements for optical system structural parameters output by the large language model, optical design formulas, and prompts and constraints for the large language model generation process.

5. The optical system design method based on knowledge base retrieval and enhanced large model according to claim 4 is characterized in that: When constructing a design requirement-parameter list data pair, the parameter list needs to be organized according to the format specified in the corresponding optical system design requirements, and then the parameter list after adjusting the format and the corresponding optical system design requirements form a design requirement-parameter list data pair.

6. The optical system design method based on knowledge base retrieval and enhanced large model according to claim 1, characterized in that: In step 3), the large language model is the Qwen2.5-14B-Instruct model; and the knowledge base retrieval enhancement module is implemented using the RAGFlow framework or the DIFY technology.

7. The optical system design method based on knowledge base retrieval and enhanced large model according to claim 2, characterized in that: In step 4), the design requirements of the optical system to be designed are consistent with the format of the optical system design requirements in the training dataset; The optical design information includes one or more of an optical system design case, a parameter range of a surface parameter, or an optical system design rule.

8. The optical system design method based on knowledge base retrieval and enhanced large model according to claim 1, characterized in that: In step 4), the optical system structural parameters include the surface shape, thickness, material, or refractive index of each surface of the optical system.

9. An optical system design system based on knowledge base retrieval and enhanced large model for implementing the design method of claim 1, characterized in that: include: A knowledge base construction module is used to obtain and preprocess optical lens design data, vectorize and encode the preprocessed optical lens design data, and construct an optical system knowledge base; A data set construction module is used to obtain a parameter list of each surface of the optical system corresponding to the optical system design requirements, construct a design requirement-parameter list data pair, and then construct a training data set based on the design requirement-parameter list data pair; A model building module is used to train the large language model using a training data set and a low-rank adaptive method to obtain a large model of optical system design; the optical system design large model is then combined with the optical system knowledge base through a knowledge base retrieval enhancement module to obtain a knowledge base retrieval enhancement large model consisting of the knowledge base retrieval enhancement module and the optical system design large model; The design module is used to obtain the design requirements of the optical system to be designed, and input them into the knowledge base retrieval enhancement module for information retrieval to obtain a number of optical design information most relevant to the design requirements. The design requirements and optical design information are then input together into the optical system design model to obtain the optical system structural parameters that meet the design requirements of the optical system to be designed, that is, to obtain a parameter list of each surface of the optical system.

Images