Geographical problem solution generation method based on geospatial case knowledge base

Abstract

This invention provides a method for generating solutions to geographical problems based on a geospatial case knowledge base, relating to the field of geospatial intelligence technology. The method includes: acquiring a geographical problem to be solved; using a large language model of a geospatial agent, retrieving candidate geospatial cases matching the geographical problem from the geospatial case knowledge base; and generating a solution for the geographical problem based on the candidate geospatial cases. The geospatial case knowledge base is constructed based on case knowledge of geospatial cases from various industry application domains, and the case knowledge of the geospatial cases includes solution ideas and solutions provided for the geographical problem. This invention enhances the ability of geospatial agents to generate solutions to geographical problems in specific industry application domains, enabling geospatial agents to generate more reasonable solutions for geographical problems in specific industry application domains.

Description

Technical Field

[0001] This invention relates to the field of geospatial intelligence technology, and in particular to a method for generating solutions to geographical problems based on a geospatial case knowledge base. Background Technology

[0002] An Artificial Intelligence Agent (AI Agent) is a computer system capable of perceiving its environment, making autonomous decisions, and executing actions to achieve specific tasks. An AI Agent consists of a Large Language Model (LLM), tools, and memory. The LLM is the brain of the geospatial agent, responsible for autonomously considering geographical problems and providing reasonable solutions. The tools are the agent's hands and feet, responsible for the specific calculations required to implement the geographical problem solution, ultimately autonomously completing the user-specified task. During the computation process, the AI ​​Agent continuously perceives environmental changes, responds accordingly through the LLM, and continuously optimizes its strategies and dynamically adjusts its behavior through interaction with the environment to achieve the optimal solution to the problem.

[0003] Geospatial agents are a type of domain-specific intelligent agent based on general AI agent technology. They are capable of understanding geospatial problems raised by users, autonomously generating solutions to geospatial problems, and performing computations to solve geospatial problems. They represent the application of general AI agents in the vertical field of geospatial information.

[0004] Currently, geospatial agents are mainly implemented based on general-purpose large language models. Although these models have been trained on geospatial information corpora during the pre-training phase and possess a certain ability to solve geospatial problems, the general-purpose large language models upon which geospatial agents rely lack knowledge of the geospatial information domain, especially geospatial case knowledge specific to industry application areas. Therefore, when faced with geospatial problems in specific application areas, geospatial agents struggle to formulate problem-solving strategies and generate solutions. Summary of the Invention

[0005] This invention provides a method for generating solutions to geographic problems based on a geospatial case knowledge base. This method addresses the shortcomings of existing technologies, which make it difficult for geospatial agents to formulate problem-solving strategies and generate solutions when facing geospatial problems in specific application areas. This invention enhances the ability of geospatial agents to generate solutions to geographic problems in specific industry application areas, enabling them to generate more reasonable solutions for geographic problems in specific industry application areas.

[0006] This invention provides a method for generating geographical problem solutions based on a geospatial case knowledge base, comprising: Obtain the geographical problem to be solved; Using a large language model of a geospatial agent, candidate geospatial cases matching the geographic problem to be solved are retrieved from a geospatial case knowledge base. Based on the candidate geospatial cases, a solution for the geographic problem to be solved is generated. The geospatial case knowledge base is constructed based on case knowledge of geospatial cases in various industry application fields. The case knowledge of geospatial cases includes solution ideas and solutions provided for geographic problems.

[0007] According to the present invention, a method for generating solutions to geographical problems based on a geospatial case knowledge base is provided, wherein the case knowledge of the geospatial case also includes the geospatial region to which the geospatial case applies.

[0008] According to the present invention, a method for generating geographical problem solutions based on a geospatial case knowledge base is provided, wherein the case knowledge of the geospatial case includes case name, case description, case input, case output, and thought chain; The case description includes the case background, research area, case function, and applicable conditions. The applicable conditions are used to record the geographic spatial area to which the geospatial case applies. The thought chain includes descriptive information of multiple independent problem-solving steps obtained by breaking down a geographical problem, used to record the problem-solving ideas and solutions provided for the geographical problem; the descriptive information of the problem-solving steps includes functional descriptions, input data descriptions, input parameter descriptions, and output data descriptions. The input data description information includes at least one of the following: data name, data function, data format, data precision, and spatial reference information; the input parameter description information includes at least one of the following: parameter name, parameter function, data type, parameter value range, default parameter, and whether it is required.

[0009] According to the present invention, a method for generating solutions to geographical problems based on a geospatial case knowledge base includes retrieving candidate geospatial cases matching the geographical problem to be solved from the geospatial case knowledge base using a large language model of a geospatial agent. Construct a first suggestion term for geospatial case retrieval; the first suggestion term contains the geospatial problem to be solved. Using the large language model of the geospatial agent, based on the geospatial basic knowledge base and the geospatial domain knowledge base, the intent understanding of the geospatial problem to be solved contained in the first prompt word is performed to obtain the intent understanding result of the geospatial problem to be solved, and based on the intent understanding result, candidate geospatial cases matching the geospatial problem to be solved are retrieved in the geospatial case knowledge base.

[0010] According to the present invention, a method for generating geographical problem solutions based on a geospatial case knowledge base is provided, wherein the geospatial domain knowledge base is constructed in the following manner: Obtain original text materials from various industry application areas; the original text materials include at least one of technical reports, technical standards, and industry reference materials. Based on the first law of geography, the second law of geography, and the geographic coordinate system, the original text materials are organized into a geospatial domain knowledge document in text format; The geospatial domain knowledge documents are converted into geospatial domain knowledge vectors and written into a vector database to obtain a geospatial domain knowledge base.

[0011] According to the present invention, a method for generating solutions to geographical problems based on a geospatial case knowledge base, wherein generating a solution to the geographical problem to be solved based on the candidate geospatial cases includes: Construct a second cue word for generating a solution to the geographic problem; the second cue word contains the candidate geospatial cases. Logical reasoning is performed based on the unresolved geographical problem and the candidate geospatial cases contained in the second prompt word to generate a solution for the unresolved geographical problem.

[0012] The present invention also provides a geographic problem solution generation device based on a geospatial case knowledge base, comprising: The problem acquisition module is used to acquire geographical problems to be solved. The solution generation module is used to utilize the large language model of the geospatial agent to retrieve candidate geospatial cases that match the geographical problem to be solved from the geospatial case knowledge base, and generate a solution for the geographical problem to be solved based on the candidate geospatial cases; wherein, the geospatial case knowledge base is constructed based on case knowledge of geospatial cases in various industry application fields, and the case knowledge of geospatial cases includes solution ideas and solutions provided for geographical problems.

[0013] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the geographic problem solution generation method based on geospatial case knowledge base as described above.

[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for generating geographic problem solutions based on a geospatial case knowledge base as described above.

[0015] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the geographic problem solution generation method based on a geospatial case knowledge base as described above.

[0016] This invention provides a method for generating geographical problem solutions based on a geospatial case knowledge base. It constructs a geospatial case knowledge base using case knowledge from geospatial cases across various industry application domains. This case knowledge includes solution approaches and solutions for geographical problems. Utilizing a large language model of a geospatial agent, it retrieves candidate geospatial cases matching the geographical problem to be solved from the knowledge base. Based on these candidate cases, a solution for the geographical problem is generated. The geospatial case knowledge base constructed by this invention supplements the limitations of general large language models in terms of geospatial case knowledge across specific industry application domains, enhancing the ability of geospatial agents to generate solutions for geographical problems in specific industry application domains. This enables geospatial agents to generate more reasonable solutions for geographical problems in specific industry application domains. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in this 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 some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0018] Figure 1 This is one of the flowcharts illustrating the method for generating geographical problem solutions based on a geospatial case knowledge base provided in this embodiment of the invention.

[0019] Figure 2 This is a schematic diagram illustrating the structure of the geospatial knowledge base provided in an embodiment of the present invention.

[0020] Figure 3This is a schematic diagram of the construction process of the geospatial knowledge base provided in the embodiments of the present invention.

[0021] Figure 4 This is the second flowchart of the method for generating geographical problem solutions based on a geospatial case knowledge base provided in this embodiment of the invention.

[0022] Figure 5 This is a schematic diagram of the structure of the geographic problem solution generation device based on a geospatial case knowledge base provided in an embodiment of the present invention.

[0023] Figure 6 This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0025] In the description of embodiments of the present invention, the terms "comprising," "including," or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Those skilled in the art will understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0026] The key to geospatial intelligent agents lies in their ability to accurately understand users' geographical problems and generate reasonable solutions to these problems. This requires geospatial intelligent agents to possess geospatial domain knowledge and the ability to transform geospatial knowledge into solutions.

[0027] Currently, geospatial agents are mainly implemented based on general-purpose large language models. Although these models have been trained on geospatial information corpora during the pre-training stage and possess a certain ability to solve geospatial problems, existing geospatial agents still have the following issues when dealing with complex geospatial problems and industry applications: Lack of geospatial domain knowledge: Case studies refer to geospatial problems, their solutions, and problem-solving steps. The large language model upon which geospatial agents rely lacks knowledge of the geospatial information domain, particularly knowledge of geospatial cases specific to industry application domains, which are characterized by their specialized nature. When faced with geospatial problems in specific domains, this makes it difficult for geospatial agents to formulate problem-solving strategies and generate solutions.

[0028] To address the aforementioned issues, this invention constructs a geospatial case knowledge base based on case knowledge from geospatial cases across various industry application domains. This geospatial case knowledge includes solution approaches and solutions for geographical problems. Utilizing a large language model of the geospatial agent, candidate geospatial cases matching the geographical problem to be solved are retrieved from the geospatial case knowledge base. Based on these candidate geospatial cases, a solution for the geographical problem is generated. The geospatial case knowledge base constructed by this invention supplements the limitations of general large language models in terms of geospatial case knowledge across specific industry application domains, enhancing the geospatial agent's ability to generate solutions for geographical problems in specific industry application domains. This enables the geospatial agent to generate more reasonable solutions for geographical problems in specific industry application domains.

[0029] Figure 1 This is one of the flowcharts illustrating the method for generating geographical problem solutions based on a geospatial case knowledge base provided in this embodiment of the invention. (Refer to...) Figure 1 This invention provides a method for generating geographical problem solutions based on a geospatial case knowledge base. The method may specifically include the following steps: Step 101: Obtain the geographical problem to be solved.

[0030] It should be noted that the execution subject of the geographic problem solution generation method based on a geospatial case knowledge base provided in this embodiment of the invention can be an electronic device, a component in an electronic device, an integrated circuit, or a chip. The electronic device can be a mobile electronic device or a non-mobile electronic device. For example, a mobile electronic device can be a mobile phone, tablet computer, laptop computer, PDA, wearable device, ultra-mobile personal computer (UMPC), netbook, or personal digital assistant (PDA), etc., while a non-mobile electronic device can be a server, network attached storage (NAS), personal computer (PC), television (TV), ATM, or self-service machine, etc. This embodiment of the invention does not specifically limit the specific implementation of these devices. The following embodiments of the invention describe the execution subject using a server as the execution subject.

[0031] Geographical problems can refer to specific needs or questions that users need to solve in the field of geography, such as questions about the causes of geographical phenomena, the distribution patterns of geographic space, and the analysis and application of geographic data.

[0032] In this embodiment of the invention, a user can input a geographical problem that needs to be solved through a user interface. The server can receive the geographical problem input by the user and then use the large language model of the geospatial agent to perform geospatial case retrieval in the geospatial case knowledge base to generate a solution for the geographical problem.

[0033] Step 102: Using the large language model of the geospatial agent, retrieve candidate geospatial cases that match the geographical problem to be solved in the geospatial case knowledge base, and generate a solution for the geographical problem to be solved based on the candidate geospatial cases; wherein, the geospatial case knowledge base is constructed based on case knowledge of geospatial cases in various industry application fields, and the case knowledge of geospatial cases includes solution ideas and solutions provided for geographical problems.

[0034] In this embodiment of the invention, the geospatial case knowledge base can be constructed based on case knowledge of specific geospatial cases in various industry application fields. A geospatial case can be a collection of geographical problems and their solution ideas and solutions.

[0035] For example, calculating the slope of terrain data is a simple geospatial case. The geographical problem, solution approach, and solution for this case can be summarized as follows: Geographical problem: Calculate the slope value of terrain data; Solution approach: Use methods or tools for calculating terrain slope to calculate the terrain slope; Solution: Use the slope tool from the gdal library to calculate the slope of the geotiff format terrain data.

[0036] In this embodiment of the invention, geospatial cases in various industry application fields can include basic spatial calculation cases (such as terrain slope calculation cases) and geospatial application cases (such as soil property spatial inference cases). Although terrain slope calculation is supported in most GIS (Geographic Information System) software, it also needs to consider the spatial reference type of the input data, and not all GIS software can correctly handle the spatial reference type of the input data; soil property spatial inference is a complex geographical problem that requires a series of spatiotemporal calculations to obtain the final data result.

[0037] In this embodiment of the invention, the goal of the geospatial case knowledge base is to embed knowledge such as geographical problems, problem-solving logic (i.e., solution approaches), and solutions into the LLM (Local Management Model), enabling the LLM to learn problem-solving knowledge for specific geographical problems. This allows the geospatial agent to generate problem-solving logic and solutions for geographical problems. When faced with a new geographical problem, the geospatial agent can generate a solution to that problem based on the learned geospatial case knowledge.

[0038] In some embodiments, geospatial case retrieval can be used to retrieve N candidate geospatial cases that match the geographical problem to be solved based on the description of the geographical problem to be solved input by the user, using the large language model of the geospatial agent. These N candidate geospatial cases are used as a case candidate set for generating a solution to the geographical problem, and are then used as a reference for the next step of generating a solution to the geographical problem.

[0039] In some embodiments, the number N of candidate geospatial cases can be determined by the user.

[0040] It should be noted that the geospatial cases in the geospatial case knowledge base for various industry application areas are standard geospatial cases set based on industry standards. The case knowledge of standard geospatial cases includes standard solution approaches and standard solutions for standard geographical problems; the geographical problem to be solved by the user can be a new geographical problem. Therefore, this embodiment of the invention utilizes the large language model of the geospatial agent to retrieve multiple standard geospatial cases that match the new geographical problem, thereby generating a solution for the new geographical problem based on the multiple matching standard geospatial cases.

[0041] In this embodiment of the invention, the large language model of the geospatial agent can generate a solution to the geospatial problem by using the logical reasoning of the geospatial agent, based on the solution ideas and solutions of the candidate geospatial cases and the problem to be solved.

[0042] This invention constructs a geospatial case knowledge base based on case knowledge from geospatial cases in various industry application domains. This geospatial case knowledge includes solution approaches and solutions for geographical problems. Using a large language model of a geospatial agent, candidate geospatial cases matching the geographical problem to be solved are retrieved from the geospatial case knowledge base. Based on these candidate geospatial cases, a solution for the geographical problem is generated. The geospatial case knowledge base constructed by this invention supplements the general large language model's lack of geospatial case knowledge in specific industry application domains, enhancing the geospatial agent's ability to generate solutions to geographical problems in specific industry application domains, enabling the geospatial agent to generate more reasonable solutions for geographical problems in specific industry application domains.

[0043] In an optional embodiment, the case knowledge of the geospatial case also includes the geospatial region to which the geospatial case applies.

[0044] Geospatial cases possess significant geospatial characteristics, primarily manifested in geographical heterogeneity, which is a significant difference between geographical problems and other problems. General problems exhibit spatial universality, meaning that a solution applicable in one spatial region will likely also be applicable in other regions, yielding the same computational results. However, for geospatial problems, due to geographical heterogeneity, a solution applicable in one geospatial region may not be applicable in another. Therefore, it is necessary to embed geographical features into the knowledge of geospatial cases.

[0045] However, existing geospatial agents suffer from a lack of geospatial thinking: geospatial problems possess significant geospatial characteristics, including basic geospatial autocorrelation and geospatial heterogeneity. The same geographical problem exhibits different spatial characteristics and solutions in different spatial regions. For example, floods in southern and northern my country have different mechanisms and therefore different solutions. However, current large language models, trained on general corpora, lack geospatial thinking capabilities such as understanding geographical heterogeneity.

[0046] This invention, by clearly defining the applicable geospatial region for a geospatial case and embedding that region into geospatial case knowledge, enables a geospatial agent to select matching geospatial case knowledge to generate a solution to the geospatial problem based on the geospatial conditions of the problem. This avoids the generation of incorrect solutions to geospatial problems caused by matching geospatial case knowledge from different geospatial regions (e.g., matching geospatial case knowledge from the south with a geospatial problem in the north).

[0047] In some embodiments, an "Applicable Conditions" item can be added to the "Case Description" section of the geospatial case description template. The applicable geospatial area and other conditions are defined in the applicable conditions. In this way, geographical features are embedded into the geospatial case through the "Applicable Conditions". When generating solutions to geographical problems, the geospatial agent can select geospatial cases applicable to the geographical problems based on the spatial area and other conditions of the geographical problems, and avoid the error of applying case knowledge embedded in the south to geographical problems in the north.

[0048] By embedding geographic features (i.e., the geographic region to which the geographic case applies) into the case knowledge of geographic spatial cases, this embodiment of the invention can further enhance the ability of a geospatial agent to solve the same geographic problem with different geographic characteristics in different spatial regions, and avoid the generation of incorrect geographic problem solutions due to matching geographic spatial case knowledge of different geographic spatial regions.

[0049] In an optional embodiment, the case knowledge of the geospatial case includes case name, case description, case input, case output, and thought chain; wherein, the case description includes case background, study area, case function, and applicable conditions, and the applicable conditions are used to record the geospatial area to which the geospatial case applies; wherein, the thought chain includes descriptive information of multiple independent problem-solving steps obtained by decomposing the geographical problem, and is used to record the solution ideas and solutions provided for the geographical problem; the description information of the problem-solving steps includes functional description of the problem-solving steps, input data description information, input parameter description information, and output data description information; wherein, the input data description information includes at least one of data name, data function, data format, data precision, and spatial reference information; the input parameter description information includes at least one of parameter name, parameter function, data type, parameter value range, default parameter, and whether it is required.

[0050] In this embodiment of the invention, the key to case knowledge in geospatial cases is the generation of the case's chain of thought (COT). The chain of thought can decompose a complex geographical problem into N independent, executable problem-solving steps, namely steps 1 to N, so that the geospatial agent can learn the problem-solving process of geographical problems.

[0051] In some embodiments, the format of the case knowledge sample for a geospatial case may be as follows: name: describe: Function: background: Applicable conditions: enter: Input data: 1. Input data 1 Data Description 2. Input data 2 Data Description Model parameters: 1. Parameter 1 Parameter Description 2. Parameter 2 Parameter Description Output: 1. Output data 1 Data Description 2. Output data 2 Data Description Thinking Chain (COT process): 1. Step 1 (1) Functional description: (2) Input data: (3) Input parameters: (4) Output data: 2. Step 2 (1) Functional description: (2) Input data: (3) Input parameters: (4) Output data: ... N. Step N (1) Functional description: (2) Input data: (3) Input parameters: (4) Output data: In some embodiments, the case knowledge sample for a geospatial case may include: (1) Case title: Describes the name of the case. The case title can concisely summarize the main content of the case.

[0052] (2) Case description: A detailed description of the case enables geospatial agents to have a comprehensive understanding of how the case is used and the spatiotemporal conditions under which it is used.

[0053] The case description may specifically include: (a) Case background: Describe the background information of the case.

[0054] (b) Study Area: Case studies are created based on geographical issues in a specific study area. The study area describes the geographical conditions and other circumstances of that study area.

[0055] (c) Case Function: Describe the function of the case.

[0056] (d) Applicability conditions: Information such as the spatial region and applicable scenarios to which the geospatial case is applicable. The case should have generalization capability based on the study area. The applicability conditions are the generalization based on the spatial conditions of the study area, describing information such as spatial regions similar to the study area, so that the geospatial agent can transplant the case to other similar spatial regions for use.

[0057] (3) Case input.

[0058] Case input can include two parts: input data and model parameters. (a) Input Data: Describes each input data point in the case. The description of each input data point may include: data name, data function, data format, data precision, spatial reference, and other data description information.

[0059] (b) Case Parameters: Describe each parameter in the case input. Parameter description information may include: parameter name, parameter function, data type, parameter value range, default parameter, whether it is required, etc.

[0060] (4) Case Output: Information describing the final output data of the case execution. This includes: data name, data format, etc.

[0061] (5) Thinking Chain (COT process): The thinking chain can decompose a complex geographical problem into N independent executable problem-solving steps, namely step 1 to step N, and then describe each problem-solving step.

[0062] For step i, the description information may include: (a) Functional description: Describe the function of step i.

[0063] (b) Input data: Describe each input data in step i, including: data name, data function, data format, data precision, spatial reference and other data description information.

[0064] (c) Input parameters: Describe each parameter in step i, including: parameter name, parameter function, data type, parameter value range, default parameter, whether it is required, etc.

[0065] (d) Output data: Step i calculates the final output data description.

[0066] When describing the steps in a thought chain, step i can be closely connected to its upstream and downstream steps i-1 and i+1 to ensure the rigor of the logic between steps. For example, the output of step i and the input of step i+1 need to be closely connected.

[0067] In one example, the following is a case knowledge example based on terrain slope calculation.

[0068] Case Name: Terrain Slope Calculation.

[0069] Case Description: Case Background: Topographic slope calculation is an analytical calculation method based on topographic data to calculate slope; Example Functionality: Calculate the slope of the terrain based on the input GeoTIFF format terrain data, and output the slope data in GeoTIFF format. Applicable conditions: The terrain slope calculation is applicable to any spatial area.

[0070] Case input: Input data: 1.input Format: GeoTIFF; Description: Input terrain raster data.

[0071] Model parameters: 1.scale Type: number; Description: The ratio of the vertical to the horizontal direction.

[0072] 2.compute_edges Type: boolean; Description: Whether to calculate edges.

[0073] Case output: 1.output Format: GeoTIFF; Description: Completed terrain slope data.

[0074] Thinking Chain (COT process): 1. Detect the spatial reference of the input data.

[0075] (1) Function description: Detect whether the spatial reference of the input data is a geographic coordinate system.

[0076] (2) Input data: Type: GeoTIFF; Description: The original input terrain raster data.

[0077] (3) Input parameters: None.

[0078] (4) Output data: Type: boolean; Description: True — Input data spatial reference is a geographic coordinate system; False — The input data space reference is the projected coordinate system.

[0079] 2. Step 2 (1) Functional description: If the spatial reference of the input data is a projected coordinate system, skip this step; If the spatial reference of the input data is a geographic coordinate system, convert the input data to a projected coordinate system, with the coordinate system code being EPSG:3859.

[0080] (2) Input data: Type: GeoTIFF; Description: The original input terrain raster data.

[0081] (3) Input parameters: none.

[0082] (4) Output data: Type: GeoTIFF; Description: Raw input terrain raster data or terrain raster data converted to the EPSG:3859 coordinate system.

[0083] 3. Step 3 (1) Functional description: Calculate the slope data from the terrain raster data.

[0084] (2) Input data: Type: GeoTIFF; Description: Terrain raster data output in step 2.

[0085] (3) Input parameters: none.

[0086] (4) Output data: Type: GeoTIFF; Description: The calculated slope raster data.

[0087] 4. Step 4 (1) Functional description: The slope data output in step 3 is projected and transformed to the coordinate system of the original input data. If the original input data is in a projected coordinate system, skip this step and output the slope data calculated in step 3. If the original input data is in a geographic coordinate system, project and transform the slope data calculated in step 3 to the coordinate system of the original input data.

[0088] (2) Input data: Input data 1: Type: GeoTIFF; Description: Terrain raster data output in step 3.

[0089] Input data 2: Type: TEXT; Description: Spatial reference for the original input data.

[0090] (3) Input parameters: none.

[0091] (4) Output data: Type: GeoTIFF; Description: The calculated slope raster data.

[0092] In one optional embodiment, the step of using the large language model of the geospatial agent to retrieve candidate geospatial cases matching the geospatial problem to be solved in the geospatial case knowledge base includes: constructing a first prompt word for geospatial case retrieval; the first prompt word contains the geospatial problem to be solved; using the large language model of the geospatial agent, based on the geospatial basic knowledge base and the geospatial domain knowledge base, performing intent understanding on the geospatial problem to be solved contained in the first prompt word to obtain the intent understanding result of the geospatial problem to be solved, and retrieving candidate geospatial cases matching the geospatial problem to be solved in the geospatial case knowledge base based on the intent understanding result.

[0093] Figure 2 This is a schematic diagram illustrating the structure of the geospatial knowledge base provided in an embodiment of the present invention. (Refer to...) Figure 2The geospatial knowledge base of a geospatial intelligent agent can consist of three knowledge bases: a basic geospatial knowledge base, a domain geospatial knowledge base, and a case geospatial knowledge base. The function of the geospatial knowledge base is to provide the geospatial intelligent agent with detailed and sufficient geospatial knowledge, enabling the agent to generate reasonable solutions to specific geographical problems and perform computational solutions. Since the basic geospatial knowledge base has already been constructed in the general large language model, this invention only requires the construction of the domain geospatial knowledge base and the case geospatial knowledge base to supplement the general large language model's deficiencies in domain geospatial knowledge and case geospatial knowledge.

[0094] In some embodiments, the geospatial domain knowledge base is primarily constructed based on fundamental knowledge of specific industry application domains. The geospatial domain knowledge base can supplement the large language model of a geospatial agent with knowledge of relevant industry application domains, enabling the geospatial agent to fully and accurately understand the geographical problems of specific industry application domains. This allows the geospatial agent to subsequently generate more reasonable solutions to the geographical problems of specific industry application domains.

[0095] For example, specific industry application areas could include disaster monitoring, land use, and other application areas.

[0096] In some embodiments, the geospatial case knowledge base can be constructed based on case knowledge of specific geospatial cases in various industry application fields, such as fire monitoring cases in the disaster field. The case knowledge of geospatial cases can provide problem-solving approaches for specific geographical problems, enabling geospatial agents to break down a complex geographical problem into N independent problem-solving steps, thus making the geographical problem solvable.

[0097] In this embodiment of the invention, a first prompt word for geospatial case retrieval can be constructed based on the user-input geospatial problem to be solved. Then, the large language model of the geospatial agent can be used to perform intent understanding on the geospatial problem to be solved contained in the first prompt word based on the constructed geospatial basic knowledge base and geospatial domain knowledge base, to obtain the intent understanding result for the geospatial problem to be solved, and based on the intent understanding result, candidate geospatial cases matching the geospatial problem to be solved can be retrieved from the geospatial case knowledge base.

[0098] In some embodiments, the semantic similarity between the intent understanding result of the geographic problem to be solved and the case names and case descriptions of various geospatial cases in the geospatial case knowledge base can be calculated, thereby determining the matching degree between the geographic problem to be solved and the geospatial cases based on the semantic similarity. The higher the semantic similarity, the higher the matching degree; the lower the semantic similarity, the lower the matching degree.

[0099] In some embodiments, the number of candidate geospatial cases can be preset by the user, or geospatial cases with semantic similarity greater than a preset similarity threshold can be determined as candidate geospatial cases. The present invention does not impose specific limitations on this.

[0100] Figure 3 This is a schematic diagram illustrating the construction process of the geospatial knowledge base provided in an embodiment of the present invention. (Refer to...) Figure 3 In some embodiments, a geospatial knowledge base for geospatial agents can be constructed using Retrieval-Augmented Generation (RAG) technology. This involves first collecting geospatial domain materials to generate raw geospatial domain knowledge data, and then collecting geospatial case materials to generate raw geospatial case knowledge data. These raw geospatial domain knowledge data and raw geospatial case knowledge data are then input into an embeddings model to generate geospatial knowledge vector representations (including geospatial domain knowledge vectors and geospatial case knowledge vectors). These geospatial knowledge vectors are then stored in a database to construct the geospatial knowledge base (geospatial domain knowledge vectors and geospatial case knowledge vectors are used to construct the geospatial domain knowledge base). Based on this geospatial knowledge base, the geospatial problem to be solved can be fully understood, and solutions to specific industry application domain geospatial problems can be generated.

[0101] This invention, through the construction of a geospatial domain knowledge base, can supplement the shortcomings of the general large language model of geospatial agents in terms of geospatial domain knowledge. By performing intent understanding on the unresolved geographical problem contained in the first prompt word based on the geospatial basic knowledge base and the geospatial domain knowledge base, the geospatial agent can more fully understand the unresolved geographical problem. Based on the intent understanding results of the unresolved geographical problem, the geospatial agent can retrieve candidate geospatial cases matching the geographical problem from the geospatial case knowledge base, enabling it to generate more reasonable solutions to the geographical problem subsequently.

[0102] In one optional embodiment, the geospatial domain knowledge base can be constructed in the following way: acquiring original text materials from various industry application areas; the original text materials include at least one of technical reports, technical standards, and industry reference materials; organizing the original text materials into geospatial domain knowledge documents in text format based on the first law of geography, the second law of geography, and the geographic coordinate system; converting the geospatial domain knowledge documents into geospatial domain knowledge vectors and writing them into a vector database to obtain the geospatial domain knowledge base.

[0103] In this embodiment of the invention, the knowledge in the geospatial domain knowledge base can originate from original textual materials such as technical reports, policy documents, technical standards, and industry reference materials in specific industry application areas. Based on geospatial domain knowledge, the geospatial agent can learn basic industry application domain knowledge. Based on geospatial domain knowledge, the geospatial agent can accurately understand the user's geographical problems, supporting further solutions to these problems.

[0104] In this embodiment of the invention, collected and aggregated original text materials such as technical reports, policy documents, technical standards, and industry reference materials can be organized into geospatial domain knowledge documents in text format, and then the geospatial domain knowledge documents can be converted into geospatial domain knowledge vectors and written into a vector database to complete the construction of a geospatial domain knowledge base.

[0105] In some embodiments, the original textual material can be organized into a geospatial domain knowledge document in text format based on the first law of geography, the second law of geography, and the geographic coordinate system. The format of a geospatial domain document can be as follows: (1) The first law of geography: Geographical things or attributes are interconnected in spatial distribution, and the correlation between neighboring things is higher than that between things that are far apart.

[0106] (2) The Second Law of Geography: Geographical phenomena are not uniformly distributed in space, but exhibit significant differences. Spatial heterogeneity can be divided into two categories: local spatial heterogeneity and stratified spatial heterogeneity. Local spatial heterogeneity refers to the difference between the attribute values ​​of a point and the attribute values ​​of its surrounding environment, such as the phenomenon of hotspots or cold spots in a certain region. Stratified spatial heterogeneity refers to the differences in attributes between different regions, such as differences in ecological zoning or geographical classification.

[0107] (3) A geographic coordinate system is a coordinate system that uses a three-dimensional sphere to define the location of points on the Earth's surface, enabling the reference of points on the Earth's surface via latitude and longitude. A geographic coordinate system consists of three parts: an angle measurement unit, the prime meridian, and a reference ellipsoid. In a spherical system, horizontal lines are lines of equal latitude or parallels of parallel. Vertical lines are lines of equal longitude or parallels of longitude.

[0108] By constructing a geospatial domain knowledge base, this invention can supplement the general large language model's lack of geospatial domain knowledge, enabling geospatial agents to fully and accurately understand geographical problems in specific industry application domains. This allows geospatial agents to generate more reasonable solutions to geographical problems in specific industry application domains.

[0109] In one optional embodiment, generating a solution to the geographic problem to be solved based on the candidate geospatial cases includes: constructing a second prompt word for generating a solution to the geographic problem; the second prompt word contains the candidate geospatial cases; and performing logical reasoning based on the geographic problem to be solved and the candidate geospatial cases contained in the second prompt word to generate a solution to the geographic problem to be solved.

[0110] Figure 4 This is the second flowchart illustrating the method for generating geographical problem solutions based on a geospatial case knowledge base, as provided in this embodiment of the invention. (Refer to...) Figure 4 In this embodiment of the invention, a first prompt word for geospatial case retrieval can be constructed based on the user-input geospatial problem to be solved. Then, based on the constructed first prompt word, the geospatial agent's large language model is used to retrieve the geospatial case knowledge base and obtain N candidate geospatial cases that best match the geospatial problem to be solved as a geospatial case candidate set for reference in the next step of generating a solution to the geospatial problem.

[0111] In this embodiment of the invention, a set of geospatial case candidates can be populated into the prompt words used to generate solutions to the geographic problem, constructing a second prompt word. Then, the large language model of the geospatial agent can use the set of geospatial case candidates contained in the second prompt word as a reference for generating solutions to the geographic problem. The geospatial agent can generate a solution to the geographic problem to be solved by learning the solution ideas and solutions of each candidate geospatial case, based on the geographic problem to be solved and using the agent's logical reasoning.

[0112] The following describes the geographic problem solution generation apparatus based on a geospatial case knowledge base provided by the present invention. The geographic problem solution generation apparatus based on a geospatial case knowledge base described below and the geographic problem solution generation method based on a geospatial case knowledge base described above can be referred to in correspondence with each other.

[0113] Figure 5 This is a schematic diagram of the structure of the geographic problem solution generation device based on a geospatial case knowledge base provided in an embodiment of the present invention. (Refer to...) Figure 5 This invention provides a device for generating geographical problem solutions based on a geospatial case knowledge base. The device may specifically include the following modules: Problem Acquisition Module 510 is used to acquire geographical problems to be solved; The solution generation module 520 is used to utilize the large language model of the geospatial agent to retrieve candidate geospatial cases that match the geographical problem to be solved in the geospatial case knowledge base, and generate a solution for the geographical problem to be solved based on the candidate geospatial cases; wherein, the geospatial case knowledge base is constructed based on the case knowledge of geospatial cases in various industry application fields, and the case knowledge of geospatial cases includes the solution ideas and solutions provided for geographical problems.

[0114] In an optional embodiment, the case knowledge of the geospatial case also includes the geospatial region to which the geospatial case applies.

[0115] In one optional embodiment, the case knowledge of the geospatial case includes case name, case description, case input, case output, and thought chain; The case description includes the case background, research area, case function, and applicable conditions. The applicable conditions are used to record the geographic spatial area to which the geospatial case applies. The thought chain includes descriptive information of multiple independent problem-solving steps obtained by breaking down a geographical problem, used to record the problem-solving ideas and solutions provided for the geographical problem; the descriptive information of the problem-solving steps includes functional descriptions, input data descriptions, input parameter descriptions, and output data descriptions. The input data description information includes at least one of the following: data name, data function, data format, data precision, and spatial reference information; the input parameter description information includes at least one of the following: parameter name, parameter function, data type, parameter value range, default parameter, and whether it is required.

[0116] In one optional embodiment, the scheme generation module includes: The first prompt word construction unit is used to construct a first prompt word for geospatial case retrieval; the first prompt word contains the geographical problem to be solved. The intent understanding unit is used to utilize the large language model of the geospatial agent, based on the geospatial basic knowledge base and the geospatial domain knowledge base, to perform intent understanding on the geospatial problem to be solved contained in the first prompt word, obtain the intent understanding result of the geospatial problem to be solved, and retrieve candidate geospatial cases that match the geospatial problem to be solved in the geospatial case knowledge base based on the intent understanding result.

[0117] In one alternative embodiment, the geospatial domain knowledge base is constructed through the following modules: The material acquisition module is used to acquire original text materials from various industry application areas; the original text materials include at least one of technical reports, technical standards, and industry reference materials. The material processing module is used to process the original text materials into geospatial domain knowledge documents in text format based on the first law of geography, the second law of geography, and the geographic coordinate system. The document conversion module is used to convert the geospatial domain knowledge documents into geospatial domain knowledge vectors and write them into a vector database to obtain a geospatial domain knowledge base.

[0118] In one optional embodiment, the scheme generation module includes: The second prompt word construction unit is used to construct a second prompt word for generating a solution to a geographic problem; the second prompt word contains the candidate geospatial cases. The solution generation unit is used to perform logical reasoning based on the geographical problem to be solved and the candidate geospatial cases contained in the second prompt word to generate a solution for the geographical problem to be solved.

[0119] This invention constructs a geospatial case knowledge base based on case knowledge from geospatial cases in various industry application domains. This geospatial case knowledge includes solution approaches and solutions for geographical problems. Using a large language model of a geospatial agent, candidate geospatial cases matching the geographical problem to be solved are retrieved from the geospatial case knowledge base. Based on these candidate geospatial cases, a solution for the geographical problem is generated. The geospatial case knowledge base constructed by this invention supplements the general large language model's lack of geospatial case knowledge in specific industry application domains, enhancing the geospatial agent's ability to generate solutions to geographical problems in specific industry application domains, enabling the geospatial agent to generate more reasonable solutions for geographical problems in specific industry application domains.

[0120] Figure 6 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 6As shown, the electronic device may include a processor 610, a communications interface 620, a memory 630, and a communication bus 640, wherein the processor 610, communications interface 620, and memory 630 communicate with each other via the communication bus 640. The processor 610 can call logical instructions in the memory 630 to execute a method for generating geographical problem solutions based on a geospatial case knowledge base. This method includes: acquiring a geographical problem to be solved; using a large language model of a geospatial agent, retrieving candidate geospatial cases matching the geographical problem to be solved from the geospatial case knowledge base; and generating a solution for the geographical problem to be solved based on the candidate geospatial cases. The geospatial case knowledge base is constructed based on case knowledge of geospatial cases in various industry application fields, and the case knowledge of the geospatial cases includes solution ideas and solutions provided for geographical problems.

[0121] Furthermore, the logical instructions in the aforementioned memory 630 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, 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 device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present 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.

[0122] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the geographic problem solution generation method based on a geospatial case knowledge base provided by the above methods. The method includes: acquiring a geographic problem to be solved; using a large language model of a geospatial agent, retrieving candidate geospatial cases matching the geographic problem to be solved in the geospatial case knowledge base, and generating a solution for the geographic problem to be solved based on the candidate geospatial cases; wherein the geospatial case knowledge base is constructed based on case knowledge of geospatial cases in various industry application fields, and the case knowledge of the geospatial cases includes solution ideas and solutions provided for geographic problems.

[0123] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements a method for generating geographical problem solutions based on a geospatial case knowledge base, as provided by the methods described above. This method includes: acquiring a geographical problem to be solved; using a large language model of a geospatial agent, retrieving candidate geospatial cases matching the geographical problem to be solved from the geospatial case knowledge base; and generating a solution for the geographical problem to be solved based on the candidate geospatial cases. The geospatial case knowledge base is constructed based on case knowledge of geospatial cases in various industry application fields, and the case knowledge of the geospatial cases includes solution ideas and solutions provided for geographical problems.

[0124] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0125] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0126] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for generating solutions to geographical problems based on a geospatial case knowledge base, characterized in that, include: Obtain the geographical problem to be solved; Using a large language model of a geospatial agent, candidate geospatial cases matching the geographic problem to be solved are retrieved from a geospatial case knowledge base. Based on the candidate geospatial cases, a solution for the geographic problem to be solved is generated. The geospatial case knowledge base is constructed based on case knowledge of geospatial cases in various industry application fields. The case knowledge of geospatial cases includes solution ideas and solutions provided for geographic problems.

2. The method for generating geographical problem solutions based on a geospatial case knowledge base according to claim 1, characterized in that, The case knowledge of the geospatial case also includes the geospatial region to which the geospatial case applies.

3. The method for generating geographical problem solutions based on a geospatial case knowledge base according to claim 1 or 2, characterized in that, The case knowledge of the geospatial case includes case name, case description, case input, case output, and thought chain; The case description includes the case background, research area, case function, and applicable conditions. The applicable conditions are used to record the geographic spatial area to which the geospatial case applies. The thought chain includes descriptive information of multiple independent problem-solving steps obtained by breaking down a geographical problem, used to record the problem-solving ideas and solutions provided for the geographical problem; the descriptive information of the problem-solving steps includes functional descriptions, input data descriptions, input parameter descriptions, and output data descriptions. The input data description information includes at least one of the following: data name, data function, data format, data precision, and spatial reference information; the input parameter description information includes at least one of the following: parameter name, parameter function, data type, parameter value range, default parameter, and whether it is required.

4. The method for generating geographical problem solutions based on a geospatial case knowledge base according to claim 1, characterized in that, The process of utilizing a large language model of a geospatial agent to retrieve candidate geospatial cases matching the geographical problem to be solved from a geospatial case knowledge base includes: Construct a first suggestion term for geospatial case retrieval; the first suggestion term contains the geospatial problem to be solved. Using the large language model of the geospatial agent, based on the geospatial basic knowledge base and the geospatial domain knowledge base, the intent understanding of the geospatial problem to be solved contained in the first prompt word is performed to obtain the intent understanding result of the geospatial problem to be solved, and based on the intent understanding result, candidate geospatial cases matching the geospatial problem to be solved are retrieved in the geospatial case knowledge base.

5. The method for generating geographical problem solutions based on a geospatial case knowledge base according to claim 4, characterized in that, The geospatial domain knowledge base is constructed in the following ways: Obtain original text materials from various industry application areas; the original text materials include at least one of technical reports, technical standards, and industry reference materials. Based on the first law of geography, the second law of geography, and the geographic coordinate system, the original text materials are organized into a geospatial domain knowledge document in text format; The geospatial domain knowledge documents are converted into geospatial domain knowledge vectors and written into a vector database to obtain a geospatial domain knowledge base.

6. The method for generating geographical problem solutions based on a geospatial case knowledge base according to claim 1, characterized in that, The step of generating a solution to the geographical problem to be solved based on the candidate geospatial cases includes: Construct a second cue word for generating a solution to the geographic problem; the second cue word contains the candidate geospatial cases. Logical reasoning is performed based on the unresolved geographical problem and the candidate geospatial cases contained in the second prompt word to generate a solution for the unresolved geographical problem.

7. A device for generating geographical problem solutions based on a geospatial case knowledge base, characterized in that, include: The problem acquisition module is used to acquire geographical problems to be solved. The solution generation module is used to utilize the large language model of the geospatial agent to retrieve candidate geospatial cases that match the geographical problem to be solved from the geospatial case knowledge base, and generate a solution for the geographical problem to be solved based on the candidate geospatial cases; wherein, the geospatial case knowledge base is constructed based on case knowledge of geospatial cases in various industry application fields, and the case knowledge of geospatial cases includes solution ideas and solutions provided for geographical problems.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the method for generating geographical problem solutions based on a geospatial case knowledge base as described in any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method for generating geographical problem solutions based on a geospatial case knowledge base as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method for generating geographical problem solutions based on a geospatial case knowledge base as described in any one of claims 1 to 6.

Images