Remote sensing science and technology knowledge service hotspot analysis method based on knowledge graph

By constructing a remote sensing knowledge graph, using crawler technology and large language models for data processing and analysis, data management problems caused by the complexity of remote sensing technology are solved, systematic sorting of knowledge and hot spot discovery are realized, and scientific research efficiency and academic exchanges are improved.

CN119090011APending Publication Date: 2024-12-06NANJING UNIV
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202411118609.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-15
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

The complexity of remote sensing technology leads to cumbersome data processing and analysis processes, lack of effective knowledge graph construction and management methods, making it difficult for industry personnel to quickly screen out suitable solutions.

Method used

Using the hot spot analysis method of remote sensing technology knowledge service based on knowledge graph, remote sensing data is obtained through crawling technology, classification and storage is performed, text classification and ontology design is used for large language models, knowledge graphs for remote sensing technology knowledge services are constructed, and knowledge graphs are oriented to remote sensing technology knowledge services are visually displayed. .

Benefits of technology

The systematized and standardized processing of knowledge in the field of remote sensing has been realized, the intelligence level of knowledge processing has been improved, the manual burden has been reduced, the dynamic update and security of data has been supported, and scientific researchers have been promoted to obtain information and exchange, and discover research hotspots and technologies. trend.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119090011A_ABST
    Figure CN119090011A_ABST
Patent Text Reader

Abstract

The invention discloses a knowledge graph-based remote sensing science and technology knowledge service hotspot analysis method, and relates to the technical field of remote sensing, and the method comprises the steps: obtaining remote sensing data from a database through employing a crawler technology, and storing the obtained remote sensing data; classifying and summarizing the stored remote sensing data, and constructing a knowledge graph oriented to the remote sensing science and technology knowledge service based on the classified and summarized remote sensing data; and based on the knowledge graph oriented to the remote sensing science and technology knowledge service, querying remote sensing data by using a statement query technology, and performing hotspot analysis on the remote sensing data. According to the method, extensive and deep knowledge acquisition in the remote sensing field is realized, and the timeliness and comprehensiveness of data are ensured; meanwhile, a large language model is utilized to intelligently analyze the thesis abstract, accurate and efficient domain classification is achieved, the intelligence level of knowledge processing is greatly improved, the labor burden is relieved, subjective errors are reduced, and a solid foundation is laid for subsequent knowledge graph construction.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of remote sensing technology, and in particular to a remote sensing science and technology knowledge service hotspot analysis method based on a knowledge graph. Background Art

[0002] Remote sensing is a non-contact, long-distance detection technology that has played an important role in agriculture, forestry, oceanography, meteorology, hydrology, environmental protection and other fields. The knowledge graph is essentially a semantic network that uses a directed graph structure to organize knowledge, where nodes represent entities or concepts and edges represent various semantic relationships between entities or concepts. It currently plays a key role in intelligent question-answering systems, intelligence analysis and other fields. As a machine learning model with a large number of parameters and a highly complex computing architecture, the large language model can handle more complex tasks and data, and demonstrates excellent expression and prediction performance.

[0003] With the rapid growth of remote sensing data, there is a lack of effective means for organizing and managing data due to the comprehensive complexity of remote sensing technology. With the help of knowledge graph technology, it is becoming increasingly important to classify and store relevant knowledge data in the remote sensing field, and to manage and analyze remote sensing data based on visualization, thereby serving remote sensing industry personnel.

[0004] At present, the research on remote sensing-related knowledge graphs is developing slowly. Although some studies have tried to assist analysis and decision-making by introducing standardized knowledge or constructing specialized knowledge graphs; or by integrating deep semantic segmentation networks with remote sensing data to explore knowledge application hotspots in the field of remote sensing; there are also works focusing on integrating massive remote sensing images with diverse unstructured data (such as information, text) through semantic association technology, and then extracting and constructing remote sensing knowledge graphs that meet the needs of data mining and remote sensing services. Due to the flexible and changeable application scenarios of remote sensing technology, the application directions of the results are innumerable. As satellites develop towards large scale, multi-spectral, high resolution, and minute levels, their application scenarios will become more and more extensive. However, because remote sensing technology is complex and comprehensive, the data processing, interpretation, and analysis processes involved are often complicated and involve multidisciplinary knowledge and skills, which makes application personnel in the industry often face a dilemma when faced with diverse remote sensing data and application needs. It is difficult to quickly and accurately select the solution that best suits the current mission scenario.

[0005] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0006] In response to the problems in the related technologies, the present invention proposes a remote sensing science and technology knowledge service hotspot analysis method based on knowledge graph to overcome the above-mentioned technical problems existing in the existing related technologies.

[0007] To this end, the specific technical solution adopted by the present invention is as follows: A remote sensing science and technology knowledge service hotspot analysis method based on knowledge graph, the method comprises the following steps: S1. Use crawler technology to obtain remote sensing data from the database and store the obtained remote sensing data; S2. Classify and summarize the stored remote sensing data, and build a knowledge graph for remote sensing science and technology knowledge services based on the classified and summarized remote sensing data; S3. Based on the knowledge graph for remote sensing scientific and technological knowledge services, use statement query technology to query remote sensing data and perform hot spot analysis on remote sensing data.

[0008] Furthermore, using crawler technology to obtain remote sensing data from a database and storing the obtained remote sensing data includes the following steps: S11. Select a database according to task requirements and use crawler technology to obtain remote sensing data from the database; S12, preprocessing the acquired remote sensing data, and classifying the sources of the preprocessed remote sensing data; S13. The remote sensing data after source classification is stored according to a storage format.

[0009] Furthermore, the storage format is a comma separated value file format.

[0010] Furthermore, the stored remote sensing data is classified and summarized, and a knowledge graph for remote sensing technology knowledge services is constructed based on the classified and summarized remote sensing data, including the following steps: S21, using text classification technology to perform text classification on the stored remote sensing data to obtain text classification results; S22. Use ontology building tools to design the ontology of text classification results, classify and summarize remote sensing data of different categories according to the ontology structure, and establish a remote sensing science and technology knowledge service platform; S23. Import the remote sensing science and technology knowledge service platform into the graph database, construct a knowledge graph for remote sensing science and technology knowledge services, and perform a visual display of remote sensing data.

[0011] Furthermore, using text classification technology to classify the stored remote sensing data, obtaining the text classification results includes the following steps: S211, analyzing the stored remote sensing data to obtain corpus categories of the remote sensing data, the corpus categories including Chinese categories and English categories; S212, determining a text classification field category according to the remote sensing data corpus category, where the text classification field category includes a Chinese classification field category and an English classification field category; S213, determining the corpus content in the remote sensing data, and based on the determination result and in combination with the corresponding text classification field category, performing text classification on the remote sensing data using a text classification algorithm to obtain a text classification result.

[0012] Further, the corpus content in the remote sensing data is determined, and based on the determination result and the corresponding text classification field category, the remote sensing data is classified using a text classification algorithm to obtain the text classification result, including the following steps: If the remote sensing data belongs to the Chinese category, the remote sensing data is classified using the knowledge enhancement model in combination with the Chinese classification domain category to obtain the text classification result; If the remote sensing data belongs to the English category, the remote sensing data is classified using a general language model in combination with the English classification domain category to obtain the text classification result.

[0013] Furthermore, if the remote sensing data belongs to Chinese paper data, the remote sensing data is classified into text using the knowledge enhancement model in combination with the Chinese classification field category, and the text classification result is obtained by the following steps: Based on the fact that remote sensing data belongs to the category of Chinese paper data, the knowledge enhancement model is integrated into the language chain framework; Set prompt words according to the Chinese classification field category; Use the language chain toolkit to read the stored remote sensing data and perform text segmentation on the remote sensing data; Combined with the prompt words and the knowledge enhancement model, the remote sensing data after text segmentation is classified to obtain the text classification results.

[0014] Furthermore, the ontology construction tool is used to design the ontology of the text classification results, and the remote sensing data of different categories are classified and summarized according to the ontology structure. The establishment of the remote sensing science and technology knowledge service platform includes the following steps: S221. According to the text classification results, select the remote sensing ontology, and design the language class diagram of remote sensing scientific and technological knowledge based on the remote sensing ontology; S222. Based on the designed language class diagram, create a classification page in the semantic wiki platform, and create the attributes of each remote sensing ontology through the classification page; S223. Combining the language class diagram with the remote sensing ontology attributes, using the ontology construction tool to construct the remote sensing ontology network language diagram; S224, importing the remote sensing ontology network language graph into the graph database, and defining the category and corresponding attributes of each remote sensing ontology; S225. Classify and summarize the defined remote sensing ontology, import the classified and summarized remote sensing ontology into the semantic wiki platform using the client library, and establish a remote sensing science and technology knowledge service platform.

[0015] Furthermore, the remote sensing science and technology knowledge service platform is imported into the graph database, a knowledge graph for remote sensing science and technology knowledge services is constructed, and a visualization of remote sensing data is performed, including the following steps: S231. Based on the remote sensing science and technology knowledge service platform, convert remote sensing data into import format; S232. Using the plug-in in the graph database, the converted remote sensing data is imported into the graph database to construct a knowledge graph for remote sensing science and technology knowledge services; S233. Use the built-in functions of the graph database to query the knowledge graph and visualize the remote sensing data based on the query results.

[0016] Furthermore, the import format is a resource description framework format.

[0017] The beneficial effects of the present invention are: 1. By integrating crawler technology, the present invention realizes extensive and in-depth knowledge acquisition in the field of remote sensing, ensuring the timeliness and comprehensiveness of the data; at the same time, it uses a large language model to perform intelligent analysis on paper abstracts, realizes accurate and efficient field classification, and greatly improves the intelligent level of knowledge processing. It not only reduces the manual burden, but also reduces subjective errors, laying a solid foundation for the subsequent construction of knowledge graphs.

[0018] 2. Through ontology definition, the present invention systematically sorts out and standardizes remote sensing scientific and technological knowledge, and constructs a knowledge graph with clear structure and clear relationships, thereby intuitively displaying the knowledge structure and association relationships in the remote sensing field in a graphical way, allowing users to quickly grasp the knowledge context and hot trends; in addition, through the application of Neo4j graph database, centralized management and efficient maintenance of remote sensing scientific and technological knowledge are realized, supporting dynamic updating, real-time synchronization and rich management tools of data, ensuring the security and consistency of data.

[0019] 3. Through the open platform interface and rich knowledge resources, the present invention allows researchers to easily obtain the required information and exchange academic views, thereby promoting the sustainable development and innovation of remote sensing technology; at the same time, based on the in-depth analysis and data mining functions of the knowledge graph, it can discover potential research hotspots and technological trends in the field of remote sensing, and provide strong decision-making support for researchers, which not only improves the efficiency and quality of scientific research, but also promotes academic exchanges and cooperation, injecting new vitality into the development of the field of remote sensing technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0021] Figure 1 is a flow chart of a remote sensing science and technology knowledge service hotspot analysis method based on a knowledge graph according to an embodiment of the present invention; Figure 2 This is an overall flow chart of the research on a hotspot analysis method for remote sensing science and technology knowledge services based on a knowledge graph according to an embodiment of the present invention; Figure 3 is a data acquisition flow chart of a remote sensing science and technology knowledge service hotspot analysis method based on a knowledge graph according to an embodiment of the present invention; Figure 4 It is a UML class diagram of a remote sensing science and technology knowledge service and a hot spot analysis method based on a knowledge graph according to an embodiment of the present invention; Figure 5 is an OWL graph of a hotspot analysis method for remote sensing scientific and technological knowledge services based on a knowledge graph according to an embodiment of the present invention; Figure 6 It is a knowledge graph (part) for remote sensing scientific and technological knowledge services and a corresponding entity node and relationship node diagram displayed on a Neo4j platform in a hotspot analysis method for remote sensing scientific and technological knowledge services based on a knowledge graph according to an embodiment of the present invention; Figure 7 It is a schematic diagram of an overview of the relationship graph between remote sensing papers and remote sensing fields and the relationship graph between each remote sensing field and remote sensing papers in the hot spot analysis method of remote sensing science and technology knowledge service based on knowledge graph according to an embodiment of the present invention. DETAILED DESCRIPTION

[0022] To further illustrate each embodiment, the present invention provides drawings, which are part of the disclosure of the present invention and are mainly used to illustrate the embodiments. They can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these contents, ordinary technicians in this field should be able to understand other possible implementation methods and advantages of the present invention.

[0023] According to an embodiment of the present invention, a remote sensing science and technology knowledge service hotspot analysis method based on knowledge graph is provided.

[0024] The present invention is further described with reference to the accompanying drawings and specific embodiments. Figure 1-Figure 7 As shown, according to the remote sensing science and technology knowledge service hotspot analysis method based on the knowledge graph according to an embodiment of the present invention, the method includes the following steps: S1. Use crawler technology to obtain remote sensing data from the database and store the obtained remote sensing data.

[0025] Specifically, using crawler technology to obtain remote sensing data from a database and storing the obtained remote sensing data includes the following steps: S11. Select a database based on task requirements and use crawler technology to obtain remote sensing data from the database.

[0026] It should be noted that the database is an open document database such as CNKI, which is used as the main data source. In the web page of the database platform, the built-in retrieval system is used to search for documents with the subject word of remote sensing. The document data is obtained through the Octopus crawler tool. The crawled data of remote sensing-themed papers include fields such as title, author, source, keywords, abstract, and publication date.

[0027] S12. Preprocess the acquired remote sensing data, and classify the sources of the preprocessed remote sensing data.

[0028] It should be noted that after crawling the required remote sensing data, that is, remote sensing paper data, these data need to be further preprocessed; first, the paper data crawled from these open source literature data platforms are cleaned, duplicate information in the title is removed, and then the paper data that is missing a certain field is deleted, and the date format of the paper publication date is unified; then, according to whether the source of the paper is from a remote sensing journal, the remote sensing papers are divided into two categories: remote sensing journal paper data and other remote sensing paper data.

[0029] In addition, journal information in the field of remote sensing is collected through the journal classification table, and domestic and international research hotspots in the field of remote sensing are summarized in the journals; at the same time, information profiles of some well-known scholars in this field are obtained from domestic and foreign remote sensing research institutions and colleges.

[0030] S13. The remote sensing data after source classification is stored according to a storage format.

[0031] Specifically, the storage format is a comma-separated value file format.

[0032] It should be noted that the comma-separated value file format is the csv (Comma-Separated Values) format.

[0033] S2. Classify and summarize the stored remote sensing data, and build a knowledge graph for remote sensing science and technology knowledge services based on the classified and summarized remote sensing data.

[0034] Specifically, the stored remote sensing data is classified and summarized, and a knowledge graph for remote sensing scientific and technological knowledge services is constructed based on the classified and summarized remote sensing data, including the following steps: S21. Use text classification technology to perform text classification on the stored remote sensing data to obtain text classification results.

[0035] It should be noted that text classification technology is a basic task in natural language processing. It can analyze the content and meaning of text and classify text into predefined labels that best match the text content. Traditional text classification methods generally require two steps: feature extraction and classifier selection, which is time-consuming and generally ineffective. Therefore, we choose to call the interface of the large language model for text classification.

[0036] Specifically, using text classification technology to classify the stored remote sensing data, obtaining the text classification results includes the following steps: S211, analyzing the stored remote sensing data to obtain corpus categories of the remote sensing data, the corpus categories including Chinese categories and English categories.

[0037] It should be noted that the choice of corpus text has a huge impact on the effect of the text classification model, and the abstract field of the paper data can highly summarize the content of the paper and extract information. Therefore, in the text classification task, the content of the paper abstract field is selected as the reference text for paper classification.

[0038] S212. Determine text classification domain categories according to the remote sensing data corpus categories. The text classification domain categories include Chinese classification domain categories and English classification domain categories.

[0039] It should be noted that before carrying out the text classification task, the field category of text classification must be determined; for the Chinese classification field category, that is, the field of Chinese remote sensing papers, it is mainly divided into seven fields: "Application of remote sensing technology in surveying and mapping", "The role of remote sensing technology in the field of national land space planning", "Application of remote sensing technology in physical geography", "Application of remote sensing in environmental resources and agriculture", "Remote sensing image processing and information measurement", "Application of remote sensing technology in industrial engineering" and "Application of remote sensing technology in people's livelihood". For the English classification field category, that is, the field of English remote sensing papers, it is mainly divided into five fields: "Unmanned Aerial Vehicle (UAV) and Satellite Imagery for Mapping", "Vegetation Decline and Geological Changes", "Dependency and Pressures on Water Resources", "Remote Sensing Image Interpretation and Technology" and "Spatiotemporal Characteristics and City Planning".

[0040] S213, determining the corpus content in the remote sensing data, and based on the determination result and in combination with the corresponding text classification field category, performing text classification on the remote sensing data using a text classification algorithm to obtain a text classification result.

[0041] It should be noted that the text classification algorithm includes a knowledge enhancement model and a general language model; the knowledge enhancement model is the ERNIE (Enhanced Representation through Knowledge Integration)-Bot model, and the general language model is the GLM (General Language Model)-4 model; in comparison, ERNIE-Bot has stronger Chinese processing capabilities and faster response speed, while GLM-4 has a more convenient deployment process and better analysis capabilities for English texts.

[0042] Therefore, in the work of text classification based on the abstract field of the obtained paper data, for the larger number of Chinese paper data, we chose to use the ERNIE-Bot model to complete the task; for the English text paper data, we chose to perform text classification based on the GLM-4 model; in addition, in the specific text classification work, it is necessary to set prompt words for the text classification task through the prompt engineering task to guide the LLM (Large Language Model) to perform text classification in specific fields; among them, the prompt words adopt strategies such as giving clear instructions, giving examples of text classification results for reference by the large language model, thinking and outputting results in two steps, and determining the output format.

[0043] Specifically, judging the corpus content in the remote sensing data, and based on the judgment result and the corresponding text classification field category, using the text classification algorithm to classify the remote sensing data, and obtaining the text classification result includes the following steps: If the remote sensing data belongs to the Chinese category, the remote sensing data is classified into text categories using the knowledge enhancement model in combination with the Chinese classification domain category to obtain the text classification results.

[0044] Specifically, if the remote sensing data contains Chinese paper data, the remote sensing data is classified by using the knowledge enhancement model in combination with the Chinese classification field category, and the text classification result is obtained by the following steps: Based on the fact that remote sensing data belongs to the category of Chinese paper data, the knowledge enhancement model is integrated into the language chain framework; Set prompt words according to the Chinese classification field category; Use the language chain toolkit to read the stored remote sensing data and perform text segmentation on the remote sensing data; Combined with the prompt words and the knowledge enhancement model, the remote sensing data after text segmentation is classified to obtain the text classification results.

[0045] It should be noted that for the text classification of Chinese abstract corpus, it is first necessary to connect the ERNIE-Bot model to the LangChain framework, explain the seven categories of Chinese text classification in the prompt words, and give examples of the content of Chinese text abstracts and the process of thinking about their classification fields, to help the ERNIE-Bot large language model think about the results of the text classification task, and then use the LangcChain toolkit to read the paper abstract field in the csv file of the remote sensing paper data, and solve the word limit of the input text by splitting the text paragraphs, and finally obtain the classification field results of each paper data in turn.

[0046] If the remote sensing data belongs to the English category, the remote sensing data is classified using a general language model in combination with the English classification domain category to obtain the text classification result.

[0047] It should be noted that for the text classification of English abstract corpus, the corresponding prompt words are set based on the zhipu toolkit that comes with GLM-4 to read the paper abstract field of the csv file of the remote sensing paper data, and then the text classification results are output and saved in sequence.

[0048] S22. Use ontology building tools to design the ontology of text classification results, classify and summarize remote sensing data of different categories according to the ontology structure, and establish a remote sensing science and technology knowledge service platform.

[0049] It should be noted that after completing the text classification, all fields of the literature data in the remote sensing field can be obtained, and the ontology design is carried out based on these fields to determine the type of ontology and the corresponding attributes.

[0050] Specifically, the ontology construction tool is used to design the ontology of the text classification results, and the remote sensing data of different categories are classified and summarized according to the ontology structure. The establishment of the remote sensing science and technology knowledge service platform includes the following steps: S221. According to the text classification results, select the remote sensing ontology, and design the language class diagram of remote sensing scientific and technological knowledge based on the remote sensing ontology.

[0051] It should be explained that the remote sensing ontology is the remote sensing entity. According to the fields of the remote sensing paper data, four types of entities can be selected, namely remote sensing papers, remote sensing journals, people, and fields. Remote sensing papers include two sub-entities, namely remote sensing journal papers and other remote sensing papers. According to these settings, the language class diagram of remote sensing scientific and technological knowledge, namely the UML (Unified Modeling Language) class diagram, is designed.

[0052] S222. Based on the designed language class diagram, create classification pages in the semantic wiki platform, and create attributes of each remote sensing ontology through the classification pages.

[0053] It should be noted that, according to the designed UML class diagram, a classification page can be created in the form of the semantic wiki platform, namely SMW (Semantic Media Wiki), including scholars, fields, remote sensing journals, remote sensing papers, and the subordinate ontology of remote sensing papers, remote sensing journal papers and other journal papers; the Schema definition is completed in the platform, and then attributes are created for each ontology; the attributes of scholars include name and description, and the attributes of fields include name and description; the attributes of remote sensing journals include journal name, editor-in-chief, place of publication, ISSN, journal language, introduction, and website. The specific attribute definition format is shown in Table 1; the attributes of remote sensing papers include title, field, publication date, abstract, keywords, and author. The attributes of remote sensing journal papers also include publishing journals, and the attributes of other remote sensing papers also include sources. The specific attribute definition formats of remote sensing journal papers and other journal remote sensing papers are shown in Tables 2 and 3, respectively.

[0054] Table 1 Attribute definition format of remote sensing journals Property Name type describe Name Text Official name of the journal Description Text Journal Introduction ChiefEditor Text Editor-in-Chief of the Journal ISSN Text Remote Sensing Journal Issue Number PublicationPlace Text Journal publication location JournalLanguage Text Journal language Website URL Journal website link Table 2 Attribute definition format of remote sensing journal papers Property Name type describe Name Text Title of the paper Author Text Author of the paper PublicationDate Date Paper publication date Abstract Text Abstract of the paper Keywords Text Keywords of the paper FieldInvolved Page Fields covered by the paper JournalPublished Page Remote sensing journals are published in journals Table 3 Attribute definition formats of other remote sensing papers Property Name type describe Name Text Title of the paper Author Text Author of the paper PublicationDate Date Paper publication date Abstract Text Abstract of the paper Keywords Text Keywords of the paper FieldInvolved Page Fields covered by the paper Resource Text Source of the paper S223. Combine the language class diagram with the remote sensing ontology attributes and use the ontology construction tool to construct the remote sensing ontology network language diagram.

[0055] It should be noted that, according to the UML class diagram and the ontology definition of the SMW platform, the remote sensing ontology network language diagram, namely the OWL (Web Ontology Language) diagram, can be further implemented in the ontology construction tool, namely the Protégé software; according to the OWL diagram, the knowledge in the remote sensing field can be clearly represented, including classes (concepts), attributes (relationships) and individuals (instances) and the various relationships between them; at the same time, the OWL diagram can more accurately describe the relationship between concepts and the characteristics of attributes, such as symmetry, transitivity, functionality, etc.

[0056] S224. Import the remote sensing ontology network language graph into the graph database, and define the category and corresponding attributes of each remote sensing ontology.

[0057] It should be noted that after importing the OWL graph in the Protégé software into the graph database, namely the Neo4j platform, the entity categories and corresponding attributes can be defined for the Neo4j graph data, ensuring the accuracy of mapping and integration into the corresponding entities and attributes when specific data is imported into the Neo4j graph database in batches. This helps solve the problem of heterogeneous data and enhances the information retrieval capabilities of the Neo4j platform.

[0058] S225. Classify and summarize the defined remote sensing ontology, import the classified and summarized remote sensing ontology into the semantic wiki platform using the client library, and establish a remote sensing science and technology knowledge service platform.

[0059] It should be noted that after the ontology design is completed, the data of various ontologies are classified and summarized, and the data is imported into the SMW platform in batches based on the Python client library, namely the mwclient library. In the SMW platform, not only can the remote sensing knowledge data be edited and managed more effectively, but also the page information display, information retrieval function, and related page jump function of the remote sensing knowledge entity can be realized.

[0060] SMW's page display has different page styles for different categories of information; for the physical page of remote sensing papers, you can directly obtain the title, author, publication date, abstract, involved fields, and keyword information of the paper on the page; for remote sensing journal papers, you can also obtain the remote sensing journal information where the paper is published; for other remote sensing papers, the page will display the source information of the paper; on the page of a remote sensing journal, you can directly obtain a series of information such as the journal name, place of publication, editor-in-chief, introduction, ISSN and journal website; the data of the journal website is URL data, and you can directly click the URL link on the page of the journal to jump to the external journal official website to view the relevant information of the remote sensing journal; when entering the page of a scholar in the remote sensing field, you can see the introduction of the scholar and the scholar's research field; when browsing the page of a remote sensing research field, the page displays the name and introduction information of the field.

[0061] The information retrieval function in the SMW platform mainly includes two methods: precise retrieval and screening retrieval. The information retrieval system is built for two types of data: remote sensing papers and remote sensing journals. Both the remote sensing paper information retrieval system and the remote sensing journal information retrieval system are equipped with two retrieval methods: precise retrieval and screening retrieval. Precise retrieval is to query the platform for a list of pages containing the entered search terms. If there are pages containing the search terms in the system, a list of pages containing the search terms will be returned after the query. Then click on a page listed in the list to enter the page and browse related information. If there is no page containing the search terms in the system, the query result will be empty. Screening retrieval is to perform information retrieval based on the filtering items provided in the retrieval system in order to find the required page information more accurately. For the remote sensing paper information retrieval system, the filter items are the various remote sensing research fields in Chinese and English. After clicking the field filter item, the list of remote sensing papers in that field will be displayed in the system. At the same time, you can enter the search terms to query the papers in that field that contain the search terms; for the remote sensing journal information retrieval system, the filter items are the journal languages ​​in Chinese and English. After clicking the language filter item, you can get a list of remote sensing journals in Chinese or English, and then enter the search terms to further query the required remote sensing journals.

[0062] The jump function of the associated page refers to the page jump between associated entities realized by the page-type data of the SMW platform. On the homepage of the platform, the jump of remote sensing fields, remote sensing papers, scholars and remote sensing journals is set. Clicking on the words can jump to the relevant classification interface; in a specific page in the remote sensing field, clicking on the paper, scholar and journal nodes in the bottom visualization map can jump to the corresponding entity information page; in the information page of a remote sensing paper, you can click on the "field" of the paper to jump to the information page of the field involved in the paper, and in the information page of a remote sensing journal paper, you can also jump to the page of the remote sensing journal where the paper is published by clicking on the "published journal" of the paper; for the page of a scholar, clicking on the "research field" of the scholar can jump to the information page of the research field.

[0063] S23. Import the remote sensing science and technology knowledge service platform into the graph database, construct a knowledge graph for remote sensing science and technology knowledge services, and perform a visual display of remote sensing data.

[0064] Specifically, the remote sensing science and technology knowledge service platform is imported into the graph database, a knowledge graph for remote sensing science and technology knowledge services is constructed, and a visualization of remote sensing data is performed, including the following steps: S231. Based on the remote sensing science and technology knowledge service platform, convert remote sensing data into import format; S232. Using the plug-in in the graph database, the converted remote sensing data is imported into the graph database to construct a knowledge graph for remote sensing science and technology knowledge services; S233. Use the built-in functions of the graph database to query the knowledge graph and visualize the remote sensing data based on the query results.

[0065] Specifically, the import format is the resource description framework format.

[0066] It should be noted that based on the remote sensing science and technology knowledge service platform, the data is exported to the Neo4j platform in the form of resource description framework format, namely RDF (Resource Description Framework); in the Neo4j platform, the RDF file of remote sensing science and technology knowledge data is imported based on the Neosemantics plug-in semantics.importRDF function; then Neo4j can realize the visualization of knowledge graph data based on its own predicate function, scalar function, aggregation function, string function and set function.

[0067] S3. Based on the knowledge graph for remote sensing scientific and technological knowledge services, use statement query technology to query remote sensing data and perform hot spot analysis on remote sensing data.

[0068] It should be noted that based on the knowledge graph, the statement query technology, namely the Cypher language, is used to query, edit, and manage the knowledge graph for remote sensing science and technology knowledge services in the Neo4j platform. For example, through the query of Cypher statements, a visualization diagram of the relationship between remote sensing papers and the field of "Application of remote sensing technology in physical geography" can be obtained; by clicking on these paper nodes, attribute information related to the paper can be viewed; overall, it can be seen that the application hotspots in this field are mainly the analysis of changing trends in the natural geographical environment.

[0069] In addition, based on the remote sensing technology knowledge graph, by querying the node relationship between remote sensing papers and remote sensing fields through Cypher statements, it can be seen that the two Chinese fields of "Application of remote sensing technology in surveying and mapping" and "Role of remote sensing technology in national land space planning" and the English fields of "Remote Sensing Image Interpretation and Technology" and "Spatiotemporal Characteristics and City Planning" cover the most literature, thus we can conclude that the current hot research directions in the remote sensing field.

[0070] In summary, with the help of the above technical solutions of the present invention, by integrating crawler technology, extensive and in-depth knowledge acquisition in the field of remote sensing is achieved, ensuring the timeliness and comprehensiveness of the data; at the same time, the large language model is used to perform intelligent analysis on the paper abstracts, achieving accurate and efficient field classification, greatly improving the intelligent level of knowledge processing, not only reducing the manual burden, but also reducing subjective errors, laying a solid foundation for the subsequent construction of knowledge graphs. Through the ontology definition work, the remote sensing scientific and technological knowledge is systematically sorted out and standardized, and a knowledge graph with a clear structure and clear relationships is constructed, thereby intuitively displaying the knowledge structure and association relationships in the remote sensing field in a graphical way, allowing users to quickly grasp the knowledge context and hot trends; in addition, through the application of the Neo4j graph database, centralized management and efficient maintenance of remote sensing scientific and technological knowledge are achieved, supporting dynamic updates of data, real-time synchronization, and rich management tools, ensuring data security and consistency. Through open platform interfaces and rich knowledge resources, researchers can easily obtain the information they need and exchange academic views, thereby promoting the continuous development and innovation of remote sensing technology. At the same time, based on the in-depth analysis and data mining functions of the knowledge graph, it can discover potential research hotspots and technological trends in the field of remote sensing, and provide strong decision-making support for researchers. It not only improves the efficiency and quality of scientific research, but also promotes academic exchanges and cooperation, injecting new vitality into the development of the field of remote sensing technology.

[0071] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A hotspot analysis method for remote sensing scientific and technological knowledge services based on knowledge graph, characterized in that: The method comprises the following steps: S1. Use crawler technology to obtain remote sensing data from the database and store the obtained remote sensing data; S2. Classify and summarize the stored remote sensing data, and build a knowledge graph for remote sensing science and technology knowledge services based on the classified and summarized remote sensing data; S3. Based on the knowledge graph for remote sensing scientific and technological knowledge services, use statement query technology to query remote sensing data and perform hot spot analysis on remote sensing data.

2. The remote sensing science and technology knowledge service hotspot analysis method based on knowledge graph according to claim 1 is characterized in that: The method of using crawler technology to obtain remote sensing data from a database and storing the obtained remote sensing data includes the following steps: S11. Select a database according to task requirements and use crawler technology to obtain remote sensing data from the database; S12, preprocessing the acquired remote sensing data, and classifying the sources of the preprocessed remote sensing data; S13. The remote sensing data after source classification is stored according to a storage format.

3. The remote sensing science and technology knowledge service hotspot analysis method based on knowledge graph according to claim 2 is characterized in that: The storage format is a comma separated value file format.

4. The remote sensing science and technology knowledge service hotspot analysis method based on knowledge graph according to claim 1 is characterized in that: The stored remote sensing data is classified and summarized, and a knowledge graph for remote sensing technology knowledge service is constructed based on the classified and summarized remote sensing data, including the following steps: S21, using text classification technology to perform text classification on the stored remote sensing data to obtain text classification results; S22. Use ontology building tools to design the ontology of text classification results, classify and summarize remote sensing data of different categories according to the ontology structure, and establish a remote sensing science and technology knowledge service platform; S23. Import the remote sensing science and technology knowledge service platform into the graph database, construct a knowledge graph for remote sensing science and technology knowledge services, and perform a visual display of remote sensing data.

5. The remote sensing science and technology knowledge service hotspot analysis method based on knowledge graph according to claim 4 is characterized in that: The text classification technology is used to perform text classification on the stored remote sensing data to obtain the text classification result, which includes the following steps: S211, analyzing the stored remote sensing data to obtain corpus categories of the remote sensing data, wherein the corpus categories include Chinese categories and English categories; S212, determining a text classification field category according to the remote sensing data corpus category, wherein the text classification field category includes a Chinese classification field category and an English classification field category; S213, determining the corpus content in the remote sensing data, and based on the determination result and in combination with the corresponding text classification field category, performing text classification on the remote sensing data using a text classification algorithm to obtain a text classification result.

6. The remote sensing technology knowledge service and hot spot analysis method based on knowledge graph according to claim 5 is characterized in that: The method of determining the corpus content in the remote sensing data and classifying the remote sensing data using a text classification algorithm based on the determination result in combination with the corresponding text classification field category to obtain the text classification result comprises the following steps: If the remote sensing data belongs to the Chinese category, the remote sensing data is classified using the knowledge enhancement model in combination with the Chinese classification domain category to obtain the text classification result; If the remote sensing data belongs to the English category, the remote sensing data is classified using a general language model in combination with the English classification domain category to obtain the text classification result.

7. The remote sensing science and technology knowledge service hotspot analysis method based on knowledge graph according to claim 6 is characterized in that: If the remote sensing data belongs to Chinese paper data, the remote sensing data is classified into text using the knowledge enhancement model in combination with the Chinese classification field category, and the text classification result is obtained, including the following steps: Based on the fact that remote sensing data belongs to the category of Chinese paper data, the knowledge enhancement model is integrated into the language chain framework; Set prompt words according to the Chinese classification field category; Use the language chain toolkit to read the stored remote sensing data and perform text segmentation on the remote sensing data; Combined with the prompt words and the knowledge enhancement model, the remote sensing data after text segmentation is classified to obtain the text classification results.

8. The remote sensing science and technology knowledge service hotspot analysis method based on knowledge graph according to claim 4 is characterized in that: The method of using the ontology construction tool to design the text classification results, classifying and aggregating remote sensing data of different categories according to the ontology structure, and establishing a remote sensing technology knowledge service platform includes the following steps: S221. According to the text classification results, select the remote sensing ontology, and design the language class diagram of remote sensing scientific and technological knowledge based on the remote sensing ontology; S222. Based on the designed language class diagram, create a classification page in the semantic wiki platform, and create the attributes of each remote sensing ontology through the classification page; S223. Combining the language class diagram with the remote sensing ontology attributes, using the ontology construction tool to construct the remote sensing ontology network language diagram; S224, importing the remote sensing ontology network language graph into the graph database, and defining the category and corresponding attributes of each remote sensing ontology; S225. Classify and summarize the defined remote sensing ontology, import the classified and summarized remote sensing ontology into the semantic wiki platform using the client library, and establish a remote sensing science and technology knowledge service platform.

9. The remote sensing science and technology knowledge service hotspot analysis method based on knowledge graph according to claim 4 is characterized in that: The method of importing the remote sensing science and technology knowledge service platform into the graph database, constructing a knowledge graph for remote sensing science and technology knowledge services, and performing a visual display of remote sensing data includes the following steps: S231. Based on the remote sensing science and technology knowledge service platform, convert remote sensing data into import format; S232. Using the plug-in in the graph database, the converted remote sensing data is imported into the graph database to construct a knowledge graph for remote sensing science and technology knowledge services; S233. Use the built-in functions of the graph database to query the knowledge graph and visualize the remote sensing data based on the query results.

10. The remote sensing science and technology knowledge service hotspot analysis method based on knowledge graph according to claim 9 is characterized in that: The import format is the Resource Description Framework format.

Citation Information

Patent Citations

  • Traffic information processing method based on uniform semantic comprehension

    CN104361017A

  • Text recognition method and device, electronic equipment and storage medium

    CN111488826A

  • NLP knowledge graph construction method combining information amount and BERT-BiLSTM-CRF

    CN114817454A

  • Multi-knowledge graph question and answer model training method and system and storage medium

    CN116011548A

  • Attack link generation method based on multi-source vulnerability data knowledge graph

    CN117061202A