Spatial data processing method and system based on large language model

Abstract

The invention discloses a spatial data processing method based on a large language model, which comprises the following steps of: acquiring spatial data to be processed, and constructing a basic database; analyzing the spatial data by using a pre-trained large language model to obtain a data analysis result, and associating the basic database with the data analysis result; receiving a data query instruction of a user by using the large language model, extracting a query result from the basic database based on a data analysis result, and feeding back the query result to the user; within a preset time period, analyzing all the data query instructions by using a large language model to obtain a query analysis result; and generating a query optimization strategy based on the query analysis result, wherein the query optimization strategy is used for optimizing the process of extracting the query result from the basic database by the large language model. The invention provides a spatial data processing method and system based on a large language model, which can be more convenient for a user to call in a subsequent calling process.

Description

Technical Field

[0001] This invention relates to the field of spatial data processing, specifically a method and system for processing spatial data based on a large language model. Background Technology

[0002] Spatial data, as the core of Geographic Information Systems (GIS) and related fields, is crucial for numerous applications such as land spatial planning, environmental monitoring, and resource management. Traditional spatial data processing methods rely on specialized data processing engines and software tools, which typically have high technical barriers, requiring users to possess corresponding professional knowledge and operational skills. Furthermore, existing technologies have the following shortcomings in terms of processing speed, visualization, and user interaction: High technical barriers: Currently, most spatial data processing methods require specialized spatial data engines and tools, which are often expensive, limiting the ease of use of spatial data. Moreover, specialized spatial data processing software is usually complex to operate, requiring users to undergo extensive training and practice to master, which restricts the popularization and application of spatial data processing technologies.

[0003] In existing technologies, spatial data processing mainly focuses on efficient storage of spatial data, lacking the mining of deep features of spatial data. When calling the data later, it mainly relies on location-based or type-based calls, making it difficult to retrieve spatial data that meets the user's needs. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a spatial data processing method and system based on a large language model. This method can add a second description to spatial data that meets the calling requirements by using custom tags. Then, it can combine the first description that meets the standards to generate a third description through a large language model. The third description can more deeply explore the features and interrelationships of the spatial data, thereby making it more convenient for users to call the data in subsequent processes.

[0005] To achieve the above objectives, the specific solution adopted by the present invention is as follows: A spatial data processing method based on a large language model includes the following steps: Acquire the spatial data to be processed and build a basic database; The spatial data is parsed using a pre-trained large language model to obtain data parsing results, and the basic database and the data parsing results are then associated. The system uses the large language model to receive the user's data query command, and extracts the query result from the basic database based on the data parsing result and then feeds it back to the user; Within a preset time period, the large language model is used to parse all the data query commands to obtain query parsing results; A query optimization strategy is generated based on the query parsing results. The query optimization strategy is used to optimize the process by which the large language model extracts the query results from the basic database.

[0006] Preferably, after obtaining the spatial data, the spatial data is initially optimized, and the basic database is constructed based on the initially optimized spatial data.

[0007] Preferably, during the initial optimization of the spatial data, feature labels are assigned to the spatial data. The feature labels include standard labels and custom labels. The standard labels are used to add a first description to the spatial data based on a standard specification, and the custom labels are used to add a second description to the spatial data based on a pre-created custom standard.

[0008] Preferably, when the first description and the second description are superimposed, the weight of the second description is greater than that of the first description.

[0009] Preferably, when parsing the spatial data using a pre-trained large language model, the large language model generates the data parsing result based on the first description and the second description.

[0010] Preferably, after the large language model generates the data parsing result, it generates a third description based on the data parsing result and adds it to the spatial data, thereby associating the data parsing result with the basic database.

[0011] Preferably, after the large language model extracts the query results and feeds them back to the user, it receives the user's evaluation of the query results and their satisfaction level.

[0012] Preferably, if the satisfaction level reaches a preset threshold, the data query instruction and the query result are used as positive samples; if the satisfaction level does not reach the preset threshold, the data query instruction and the query result are used as negative samples.

[0013] Preferably, when the large language model parses all the data query instructions, it assigns sample weights to the positive samples and the negative samples, and the sample weight of the positive samples is greater than the sample weight of the negative samples.

[0014] A spatial data processing system based on a large language model is provided to implement the aforementioned spatial data processing method based on a large language model. The system includes: The data acquisition module is used to acquire the spatial data; A data storage module is used to construct the basic database; The model execution module is used to run the pre-trained large language model. The human-computer interaction module is used for human-computer interaction with users.

[0015] This invention can add a second description to spatial data that meets the calling requirements by custom tags, and then combine the first description that conforms to the standard to generate a third description through a large language model. The third description can more deeply explore the features and interrelationships of the spatial data, thus making it more convenient for users to call in subsequent processes. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a structural block diagram of the system of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Please see Figure 1 and 2 , Figure 1 This is a flowchart of the method of the present invention. Figure 2 This is a structural block diagram of the system of the present invention.

[0020] A spatial data processing method based on a large language model includes the following steps: S1: Acquire the spatial data to be processed and construct a basic database. The specific form and storage format of the spatial data are mature existing technologies and will not be elaborated here.

[0021] In order to better store spatial data and facilitate subsequent processing, the spatial data is initially optimized after acquisition, and a basic database is built based on the initially optimized spatial data.

[0022] More specifically, during the initial optimization of spatial data, feature labels are assigned to the spatial data. These feature labels include standard labels and custom labels. Standard labels are used to add a primary description to the spatial data based on standardized criteria, while custom labels are used to add a secondary description based on pre-created custom criteria. Standard labels mark which standards the spatial data conforms to; these standards can be national, industry, or enterprise standards, etc. By assigning standard labels to spatial data, standardized management of the data can be achieved, preventing serious errors. Custom labels can be determined according to actual needs, such as information about the data collectors. By assigning custom labels to spatial data, more personalized features can be added, facilitating subsequent processing and making it easier to view and retrieve the spatial data.

[0023] A standard library for storing standard tags and a custom library for storing custom tags can be created separately. The required standard tags are stored in the standard library, and a first index is added to each standard tag in the standard library. The required custom tags are added to the custom library, and a second index is added to each custom tag in the custom library. When assigning feature tags to spatial data, the first and second indices can be added as feature characters to the beginning or end of the spatial data. This approach is easier to implement, especially when the character length of custom tags is long. It avoids excessively increasing the data volume of the spatial data after adding feature tags, thereby reducing the hardware cost of storing the spatial data.

[0024] Considering that custom tags better meet practical needs, when the first and second descriptions are combined, the second description has a higher weight than the first description. In subsequent processing and invocation, the second description is given priority, followed by the first description.

[0025] S2: A pre-trained large language model is used to parse spatial data, yielding parsing results, and then the basic database and the parsing results are linked. Specifically, when parsing spatial data using the pre-trained large language model, the model generates parsing results based on the first and second descriptions. The large language model primarily parses the first and second descriptions to achieve semantic understanding, without parsing the spatial data itself. Through the large language model's parsing, a deeper understanding of the first and second descriptions can be achieved, uncovering deeper connections between them for subsequent retrieval of spatial data.

[0026] After the large language model generates data parsing results, a third description is generated based on these results and added to the spatial data, thus linking the data parsing results with the basic database. By generating the third description, it can work in conjunction with the first and second descriptions to provide a more in-depth description of the spatial data. Utilizing the first, second, and third descriptions allows for a comprehensive and accurate description of the characteristics of the spatial data, facilitating subsequent processing.

[0027] For example, in two sets of spatial data, one set's second description is a house, and the other set's second description is a parking lot. The two sets of spatial data are spatially related, meaning the parking lot is adjacent to the house. The generated third description could be that users of the house can use the parking lot, or that the parking lot can be used by users of the house. It can be seen that by generating a third description, the potential relationships between different spatial data can be further explored based on the custom labels of the second description, thus better establishing the relationships between different spatial data. The third description can be generated based on data analysis results. Taking the house and parking lot as an example again, the data parsing result from the large language model could be that the distance between the house and the parking lot is less than a preset threshold, and that the house and parking lot have compatible functions. Based on this, the aforementioned third description can be generated.

[0028] S3: The large language model receives user data query commands and extracts query results from the basic database based on the data parsing results before feeding them back to the user. Specifically, after receiving the user's data query command, the large language model first parses the true semantics of the data query command, and then searches the basic database based on the true semantics, especially based on the first, second, and third descriptions. For example, if a user's data query command is "parking lot available for a certain house resident," the large language model, after understanding the data query command, determines the house and surrounding parking lots based on the second description, and then determines whether the parking lot can be used by the house resident based on the third description. After obtaining the search results, the large language model performs natural language processing on the search results to generate language that can be understood by the user and feeds it back to the user, thus realizing the response to the user's data query command. Based on the aforementioned first, second, and third descriptions, the large language model can generate results that fully satisfy the user's data query command, thereby ensuring that the user can quickly query and retrieve spatial data.

[0029] S4: Within a preset time period, use a large language model to parse all data query commands to obtain query results.

[0030] S5: Generate query optimization strategies based on query parsing results. These strategies are used to optimize the process by which the large language model extracts query results from the basic database.

[0031] After the large language model extracts the query results and returns them to the user, it receives the user's evaluation of the query results and their satisfaction level. Although the pre-trained large language model can understand the user's data query instructions and generate corresponding results, errors are still possible in practical applications. That is, there may be a discrepancy between the results returned by the large language model and the user's data query instructions, resulting in the user not obtaining the desired results. To continuously optimize the large language model, this invention introduces a real-time evaluation mechanism. While providing search results to the user, it receives the user's evaluation of their satisfaction level, thus determining whether the user is satisfied with the search process based on the evaluation satisfaction. Furthermore, if the evaluation satisfaction reaches a preset threshold, the data query instructions and results are considered positive samples; if the evaluation satisfaction does not reach the preset threshold, the data query instructions and results are considered negative samples. Positive and negative samples can be collected periodically, and these samples can be used to continuously optimize and train the large language model, thereby continuously improving its performance.

[0032] When parsing all data query commands, the large language model assigns weights to positive and negative samples, with positive samples having a greater weight than negative samples. By adjusting the weights of positive and negative samples, and ensuring that the weight of positive samples is greater than that of negative samples, the large language model can gradually align with the query process based on positive samples, thereby optimizing the large language model.

[0033] In this invention, the large language model can adopt mature existing models, and the existing models can be adaptively optimized using spatial data. For example, the Deepseek model or the Chat-GPT model can be adopted, which are mature existing technologies and will not be described in detail here.

[0034] This invention can add a second description to spatial data that meets the calling requirements by custom tags, and then combine the first description that conforms to the standard to generate a third description through a large language model. The third description can more deeply explore the features and interrelationships of the spatial data, thus making it more convenient for users to call in subsequent processes.

[0035] The present invention further provides a spatial data processing system based on a large language model, for implementing the above-mentioned spatial data processing method based on a large language model. The system includes a data acquisition module, a data storage module, a model running module, and a human-computer interaction module.

[0036] The data acquisition module is used to acquire spatial data.

[0037] A data storage module is used to build a basic database. The data storage module includes at least one memory.

[0038] The model execution module is used to run pre-trained large language models. The model execution module includes a processor.

[0039] The human-computer interaction module is used for human-computer interaction with users.

[0040] It should be understood that the processor in the embodiments of the present invention can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0041] It should also be understood that the memory in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDR SDRAM), enhanced synchronous DRAM (ESDRAM), synchronous linked DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0042] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage system such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0043] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0044] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0045] In the embodiments provided by this invention, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0046] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer system (which may be a personal computer, server, or network system, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0047] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0048] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A spatial data processing method based on a large language model, characterized in that, Includes the following steps: Acquire the spatial data to be processed and build a basic database; The spatial data is parsed using a pre-trained large language model to obtain data parsing results, and the basic database and the data parsing results are then associated. The system uses the large language model to receive the user's data query command, and extracts the query result from the basic database based on the data parsing result and then feeds it back to the user; Within a preset time period, the large language model is used to parse all the data query commands to obtain query parsing results; A query optimization strategy is generated based on the query parsing results. The query optimization strategy is used to optimize the process by which the large language model extracts the query results from the basic database.

2. The spatial data processing method based on a large language model as described in claim 1, characterized in that, After obtaining the spatial data, the spatial data is initially optimized, and the basic database is constructed based on the initially optimized spatial data.

3. The spatial data processing method based on a large language model as described in claim 2, characterized in that, During the initial optimization of the spatial data, feature labels are assigned to the spatial data. The feature labels include standard labels and custom labels. The standard labels are used to add a first description to the spatial data based on the standard specifications, and the custom labels are used to add a second description to the spatial data based on the pre-created custom standards.

4. The spatial data processing method based on a large language model as described in claim 3, characterized in that, When the first description and the second description are superimposed, the weight of the second description is greater than that of the first description.

5. The spatial data processing method based on a large language model as described in claim 3, characterized in that, When parsing the spatial data using a pre-trained large language model, the large language model generates the data parsing result based on the first description and the second description.

6. The spatial data processing method based on a large language model as described in claim 5, characterized in that, After the large language model generates the data parsing result, it generates a third description based on the data parsing result and adds it to the spatial data, thereby associating the data parsing result with the basic database.

7. The spatial data processing method based on a large language model as described in claim 1, characterized in that, After the large language model extracts the query results and returns them to the user, it receives the user's evaluation of the query results and their satisfaction level.

8. The spatial data processing method based on a large language model as described in claim 7, characterized in that, If the satisfaction rating reaches a preset threshold, the data query instruction and the query result are used as positive samples; if the satisfaction rating does not reach the preset threshold, the data query instruction and the query result are used as negative samples.

9. The spatial data processing method based on a large language model as described in claim 8, characterized in that, When the large language model parses all the data query commands, it assigns sample weights to the positive samples and the negative samples, and the sample weight of the positive samples is greater than the sample weight of the negative samples.

10. A spatial data processing system based on a large language model, characterized in that, The system is used to implement the spatial data processing method based on a large language model as described in any one of claims 1-9, the system comprising: The data acquisition module is used to acquire the spatial data; A data storage module is used to construct the basic database; The model execution module is used to run the pre-trained large language model. The human-computer interaction module is used for human-computer interaction with users.

Images