Open source information collection and integration system and method based on geographic grid coding

Abstract

The invention discloses an open source information collection and integration system and method based on geographic grid coding, and belongs to the technical field of information collection. The method comprises the following steps: obtaining a to-be-divided geographic area and to-be-judged information; the method comprises the following steps: performing geographic space grid hierarchical system division on a to-be-divided geographic region, and performing geographic grid coding on each grid unit; after all information elements in the grid are stored based on geographic grid codes, entity relation extraction is carried out to generate a triple knowledge graph; performing multi-source heterogeneous data cleaning and disambiguation and event time mapping on the geographic grid after the generation of the triple knowledge graph; and carrying out association relation judgment and credibility judgment on the to-be-judged information based on the geographic grid after the event time mapping is completed. According to the method, global subdivision grid coding is utilized, any geographic area is converted into a computable grid set, multi-source and multi-mode open source information in the area is crawled, and rapid convergence, cross verification and credible presentation of the information are achieved.

Description

Technical Field

[0001] This invention relates to an open-source information collection and integration system and method based on geographic grid coding, belonging to the field of information collection technology. Background Technology

[0002] With the rapid development of internet technology, online information is exploding, and open-source information is of great value for decision-making, risk assessment, and situational awareness in various fields. However, current open-source information collection methods often lack precise focus on specific spatial areas. When collecting information in a particular region, it is difficult to quickly and accurately define the boundaries of the search, resulting in information that is either too broad or too narrow, failing to meet actual needs. For example, when collecting city-level business information, a large amount of irrelevant information from surrounding cities or rural areas may be mixed in, or key information from some peripheral areas of the city may be missed.

[0003] Existing open-source information gathering methods often lack precise focus on specific spatial ranges. When collecting information in a certain area, it is difficult to quickly and accurately define the boundaries to be searched within a complex spatial range. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and provide an open-source information collection and integration system and method based on geographic grid coding. Based on the spatial range defined by the geographic coding grid, relevant information within the region is collected from public channels, and then summarized and refined to improve the accuracy and speed of open-source information search.

[0005] To achieve the above objectives, the present invention is implemented using the following technical solution: In a first aspect, the present invention provides an open-source information collection and integration method based on geographic grid coding, comprising: Obtain the geographical region to be divided and the information to be determined; After dividing the geographic region into a geospatial grid hierarchy, each grid unit is geocoded. After storing all information elements within the geographic grid based on the geographic grid encoding, entity relationships are extracted to generate a triplet knowledge graph. Multi-source heterogeneous data cleaning and disambiguation, as well as event time mapping, are performed on the geographic grid after the triplet knowledge graph is generated. Based on the geographic grid after the event time mapping has been completed, the correlation and credibility of the information to be judged are determined, and the open source information is collected and integrated based on the judgment results.

[0006] Furthermore, after dividing the geographic region into a geospatial grid hierarchy, each grid unit is geocoded, including: dividing the Earth's surface into multiple levels from high to low, where the longitude and latitude spans of the k-th level are: Δλ k =360 / (2^k×m) Δφ k =180 / (2^k×n) In the formula: Δλ k For longitude span, Δφ k Where k is the latitudinal span, m and n are the number of grid levels, and m and n are constants related to the grid level. For each grid cell, it is encoded according to hierarchical information, longitude index, and latitude index, as shown in the expression: ID = k × 10^(2L) + i × 10^L + j i=floor((λ-λ0) / Δλ k ) j=floor((φ-φ0) / Δφ k ) In the formula: ID is the unique code of the grid, L is the maximum number of bits for indices i and j, i is the index in the longitude direction, j is the index in the latitude direction, λ is the longitude of the grid cell center, φ is the latitude of the grid cell center, λ0 is the starting longitude, φ0 is the starting latitude, and floor is the floor function.

[0007] Furthermore, after storing all information elements within the geographic grid based on the geographic grid encoding, entity relations are extracted to generate a triplet knowledge graph, including: Using the grid ID as the key, all information elements within the grid are stored, establishing a correspondence between the grid and the information elements. The information elements include domain names, IP addresses, social media accounts, and event reports. Entity relations are extracted from text / images using a BERT-based entity recognition and relation extraction model. The extracted results are then used to generate a triplet knowledge graph containing grid IDs, topic words, and confidence scores. The confidence score is calculated using the following formula: C = α × P + β × R Where C is the confidence level, P is the accuracy of entity relation extraction, R is the ratio of the frequency of the information element appearing in the grid to the total frequency, and α and β are weighting coefficients with α+β=1.

[0008] Furthermore, the generated triplet knowledge graph mesh undergoes multi-source heterogeneous data cleaning and disambiguation, as well as event-time mapping, including: By integrating geocoding APIs from Gaode, Baidu, and Google, as well as a self-built postal code / POI knowledge graph, geocoding is performed on each API for the address information to be processed, resulting in multiple candidate coordinates. The similarity between each candidate coordinate and the corresponding address coordinates in the self-built knowledge graph is calculated, and the coordinate with the highest similarity is selected as the final coordinate of the address. A temporal convolutional network is used to predict the actual occurrence time of the event. The input of the temporal convolutional network is a feature vector containing the publication time and time descriptions in the event-related text, and the output is the time interval of the event. The predicted event time interval is mapped to the corresponding grid. The mapping rule is: if there is valid time coverage of the grid within the event occurrence time interval, then the event is mapped to that grid. The similarity calculation formula is as follows: S=1-d / D max In the formula: S is the similarity, d is the distance between the two coordinates, and D is the similarity score. max This is the maximum distance threshold that is set.

[0009] Furthermore, based on the geographic grid after the event time mapping is completed, the correlation and credibility of the information to be judged are determined, including: Calculate the temporal and geographical similarity of the information to be judged based on the geographic grid after the event time mapping has been completed; The spatiotemporal similarity of the information to be judged is calculated based on the temporal similarity and geographical similarity of the information to be judged. Calculate the degree of information association between the information to be judged based on the spatiotemporal similarity of the information to be judged; The degree of information correlation of the information to be judged is compared with a set threshold. If the degree of information correlation is greater than the set threshold, it is judged that the information has a correlation; otherwise, it is judged that there is no correlation. The credibility of the information to be judged is determined by calculating the total credibility score of the information to be judged through a three-level verification mechanism.

[0010] Furthermore, the formula for calculating the degree of information correlation is as follows: R= ×Sst+(1- )×Ss Ss=V A V B / (||V A ||×||V B ||) Sst = γ × St + (1 - γ) × Sg St=1-|t A -t B | / T max Sg=1-dAB / G max In the formula: R represents the degree of information correlation. Here, Sst represents the weighting coefficients for spatiotemporal similarity, Ss represents semantic similarity, and V represents the weighting coefficients for spatiotemporal similarity. A Let V be the semantic feature vector of information A. B Let γ be the semantic feature vector of information B, γ be the weight coefficient of temporal similarity, St be the temporal similarity, and t be the value of t. A Let t be the timestamp of information A. B Let T be the timestamp of information B. max The maximum time difference threshold is set, Sg is the geographical similarity, and d is the maximum time difference threshold. AB G represents the distance between the geographic coordinates of information A and information B. max This is the maximum geographic distance threshold that is set.

[0011] Furthermore, the three-level verification mechanism includes primary verification, secondary verification, and tertiary verification, wherein: The first-level verification includes checking whether the information appears on at least two different types of open source platforms. If it appears, it is recorded as passing the first-level verification and a first-level verification score is assigned. The secondary verification includes, based on the primary verification, analyzing whether the key elements in the information are consistent, and calculating the consistency score C of the key elements. key When C key When the value is greater than or equal to the preset threshold b, it is considered as passing the second-level verification, and a second-level verification score is assigned and the first-level verification score is accumulated. The three-level verification includes, based on passing the two-level verification, combining historical data and domain knowledge, judging whether the information conforms to logic and common sense, calculating the conformity degree F through logical reasoning and knowledge matching, and recording the passing of the three-level verification when F ≥ a preset threshold d, assigning a three-level verification score and accumulating the scores of the first and second-level verification as the total score. When the total score is greater than or equal to the preset total score threshold f, it is marked as "high confidence"; when the preset total score threshold g is less than the preset total score threshold f, it is marked as "medium confidence"; when the total score is less than the preset total score threshold g, it is marked as "low confidence", where the preset total score threshold g is less than the preset total score threshold f.

[0012] Secondly, the present invention provides an open-source information collection and integration system based on geographic grid coding, comprising: Information acquisition module: Acquires the geographical region to be divided and the information to be determined; Grid coding module: After dividing the geographic region to be divided into a geospatial grid hierarchy, each grid unit is geogrid coded. The graph generation module: After storing all information elements within the geographic grid based on the geographic grid encoding, it extracts entity relationships to generate a triplet knowledge graph; Data mapping module: performs multi-source heterogeneous data cleaning and disambiguation, as well as event time mapping, on the geographic grid after generating the triplet knowledge graph; Judgment and Integration Module: Based on the geographic grid after the event time mapping has been completed, the module performs correlation and credibility determination on the information to be judged, and completes open source information collection and integration based on the judgment results.

[0013] Thirdly, the present invention provides an open-source information collection and integration device based on geographic grid coding, including a processor and a storage medium; The storage medium is used to store instructions; The processor is configured to operate according to the instructions to perform the steps of the method according to any of the foregoing.

[0014] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any of the methods described above.

[0015] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: This solution provides a unified "spatial scale" for all diverse and inconsistent open-source information through mandatory gridding, fundamentally solving the problem of fusion of multi-source heterogeneous spatial data. Simultaneously, it uses geographic grid coding as a primary index, transforming complex spatial range queries into efficient one-dimensional coding prefix matching, making it particularly suitable for rapid filtering and aggregation of massive amounts of data—a significant advantage. More importantly, structured and standardized grid data is an ideal input for advanced data mining algorithms such as machine learning and deep learning, paving the way for more intelligent application scenarios such as spatiotemporal behavior prediction and anomaly pattern detection. Attached Figure Description

[0016] The accompanying drawings, which form part of this specification, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings: Figure 1 This is a flowchart illustrating an open-source information collection and integration method based on geographic grid coding, as provided in Embodiment 1 of the present invention. Detailed Implementation

[0017] The present invention will now be described in detail with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described herein can be combined with each other.

[0018] The following detailed description is exemplary and intended to provide further detailed explanation of the invention. Unless otherwise specified, all technical terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in this invention is for describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention.

[0019] Example 1: Please see Figure 1 To address the problems of inaccurate geospatial definition, incomplete information collection, weak integration capabilities, and low credibility in existing technologies, this solution designs an open-source information collection and integration method based on geographic grid coding. This method uses a geographic coding grid as a spatial reference framework, combines intelligent keyword generation and multi-source crawling technologies to comprehensively collect information, and then completes information integration and credibility verification through preprocessing, association modeling, and interactive verification, ultimately forming high-quality open-source information. The overall technical roadmap is described step by step as follows: (1) First, based on the geographic grid coding rules (taking the Geospatial Grid Coding Rules (GB / T40087-2021) as an example), the geographic spatial grid hierarchy system division and coding method are determined, and a unique code is generated for each grid unit.

[0020] (2) Then, construct an open-source information "grid-semantic" dual index. Using the grid ID as the key, store all information elements (domain name, IP, social account, event report, etc.) within the grid, and extract entity relationships from text / images to generate (grid ID, keyword, confidence) triples to achieve "searching for events by grid and finding grids by events".

[0021] (3) Secondly, multi-source heterogeneous data cleaning and disambiguation. Integrating the geocoding APIs of Gaode / Baidu / Google and the self-built postal code / POI knowledge graph, the problem of same name but different address and different name but same address is solved. The "publication time" and "event time" are separated, and the actual occurrence time of the event is predicted by the temporal convolutional network (TCN) and mapped to the corresponding grid.

[0022] (4) Next, construct an information association model based on spatiotemporal and semantic features. Spatiotemporal features include the timestamp and geographic coordinates of the information occurrence; semantic features include the theme, keywords, and sentiment of the event. By calculating the spatiotemporal and semantic similarity between different pieces of information, the association between the information is established.

[0023] (5) Finally, multi-source information is used for verification. Information sources are classified, a confidence assessment system is constructed, the authenticity of information is verified, and a three-level verification mechanism is set up: Level 1 verification: check whether the information appears on at least two different types of open source platforms; Level 2 verification: analyze whether the key elements in the information (such as time, place, and people) are consistent; Level 3 verification: combine historical data and domain knowledge to judge whether the information is logical and reasonable. Information that passes the three-level verification is marked as "high credibility", and information that only passes the first or second level verification is marked as "medium credibility" or "low credibility".

[0024] The following is a further detailed description of this plan: To address the current issues in open-source information collection, which uses keywords or administrative regions as retrieval units and lacks a unified spatial benchmark, leading to difficulties in aligning cross-source and cross-scale data, inconsistencies in public data formats, coordinate systems, and semantics, and the absence of automated conflict detection and confidence assessment mechanisms, this solution employs an open-source information collection and integration method based on geographic grid coding. This method aggregates, manages, analyzes, and evaluates multi-source heterogeneous data using geographic grid coding, transforming "any public information" into "event records with Cell-IDs." A three-dimensional cube model of "grid-semantics-time" is used to achieve large-scale collection, fusion, retrieval, and distribution of open-source information. Based on the above analysis, the specific steps are as follows: S1: Global Grid Subdivision and Coding Construction. Based on user needs or existing national standards, the Earth's surface or a specific region is divided into cells of equal size or different levels, and each cell is assigned a globally or regionally unique code or identifier.

[0025] S1-1: Geospatial Grid Hierarchy. The Earth's surface is divided into n levels from high to low. The higher the level, the higher the grid accuracy. The longitude span Δλ of the grid at level k (1≤k≤n) is... k and latitude span Δφ k Determine according to the following formula: Δλ k =360 / (2^k×m) Δφ k =180 / (2^k×n) Where m and n are level-related constants, set according to actual needs, and satisfying the requirement that the mesh can achieve the minimum accuracy at the highest level.

[0026] S1-2: Geographic Grid Coding. For each grid cell, its code consists of hierarchical information, a longitude index, and a latitude index. Let a grid cell be located at the k-th level, with longitude index i and latitude index j. Then the unique ID of this grid cell can be calculated using the following formula: ID = k × 10^(2L) + i × 10^L + j Where L is the maximum number of bits for indices i and j, ensuring that i and j can be fully represented. i and j are calculated as follows: i=floor((λ-λ0) / Δλ k ) j=floor((φ-φ0) / Δφ k ) Where λ is the longitude of the grid cell center, φ is the latitude of the grid cell center, λ0 is the starting longitude (e.g., -180°), φ0 is the starting latitude (e.g., -90°), and floor is the floor function.

[0027] S2: Construction of Open Source Information "Grid-Semantic" Dual Index S2-1: Using the grid ID as the key, store all information elements within the grid, such as domain names, IP addresses, social media accounts, event reports, etc., and establish the correspondence between grids and information elements.

[0028] S2-2: Entity relation extraction is performed on text / images using a BERT-based entity recognition and relation extraction model. The extracted results are used to generate a (grid ID, topic term, confidence score) triple knowledge graph. The confidence score C is calculated using the following formula: C = α × P + β × R Where P is the accuracy of entity relationship extraction, R is the ratio of the frequency of the information element appearing in the grid to the total frequency, α and β are weight coefficients, and α+β=1, which can be adjusted according to the actual data distribution and business needs. The function of "searching for events by grid and finding grids by events" is realized through this triple.

[0029] S3: Multi-source heterogeneous data cleaning, disambiguation, and event time mapping S3-1: Multi-source heterogeneous data cleaning and disambiguation. This involves integrating the geocoding APIs of Gaode, Baidu, and Google, along with a self-built postal code / POI knowledge graph. For the address information to be processed, each API is called for geocoding, resulting in multiple candidate coordinates. The similarity between each candidate coordinate and the corresponding address coordinates in the self-built knowledge graph is calculated. The formula for calculating the similarity S is: S=1-d / D max Where d is the distance between the two coordinates, which can be calculated using latitude and longitude. max This is the maximum distance threshold set. The coordinates with the highest similarity are selected as the final coordinates of the address, resolving the issues of identical names at different addresses and different names at the same address.

[0030] S3-2: Event Time Processing. The "publication time" and "event time" are separated, and a Temporal Convolutional Network (TCN) is used to predict the actual occurrence time of the event. The input to the TCN is a feature vector containing information such as the publication time and time descriptions in the event-related text; the output is the time interval of the event. The predicted event time interval is mapped to a corresponding grid. The mapping rule is: if there is valid time coverage of a grid within the event time interval, then the event is mapped to that grid.

[0031] S4: Information Association Model Construction S4-1: Spatiotemporal Similarity Calculation Let the timestamp of information A be t. A The geographic coordinates are (λ A , φ A The timestamp of information B is t. B The geographic coordinates are (λ B , φ B The formula for calculating the time similarity St is: St=1-|t A -t B | / T max Among them, T max This is the maximum time difference threshold that is set.

[0032] The formula for calculating geographical similarity Sg is: Sg=1-d AB / G max Where, d AB G represents the distance between the geographic coordinates of information A and information B. max This is the maximum geographic distance threshold that is set.

[0033] Spatiotemporal similarity Sst = γ × St + (1 - γ) × Sg, where γ is the weighting coefficient of temporal similarity, 0 < γ < 1.

[0034] S4-2: Semantic Similarity Calculation The similarity of the event's topic and keywords is calculated using word vector cosine similarity, while sentiment similarity is calculated based on the difference in sentiment scores. Let V be the semantic feature vector of information A. A The semantic feature vector of information B is V. B Then the formula for calculating semantic similarity Ss is: Ss=V A V B / (||V A ||×||V B ||) in, For vector dot product, || || represents the magnitude of the vector.

[0035] S4-3: Calculation of Information Relevance R R= ×Sst+(1-δ)×Ss, where The weighting coefficient for spatiotemporal similarity, 0 < <1. When R is greater than the set threshold, the two pieces of information are considered to be correlated.

[0036] S5: Verification through Multi-Source Information Interaction S5-1: Information Source Classification. Information sources are classified into different types, such as official media, authoritative institutions, and content generated by ordinary users, and each type is assigned a different initial credibility value.

[0037] S5-2: Confidence assessment system and three-level verification mechanism.

[0038] Level 1 Verification: Check if the information appears on at least two different types of open-source platforms. If it does, it is considered to have passed Level 1 verification and is assigned a score of 'a'.

[0039] Secondary corroboration: Building upon primary corroboration, this involves analyzing the consistency of key elements in the information (such as time, place, and people). A consistency score (C) for these key elements is calculated. key A preset threshold b is set when C key When the value is greater than or equal to the threshold b, it is recorded as passing the second-level verification, and the total score becomes a+c (c is the second-level verification score).

[0040] Level 3 verification: Building upon Level 2 verification, this step combines historical data and domain knowledge to determine if the information conforms to logic and common sense. The degree of conformity, F, is calculated through logical reasoning and knowledge matching, with a preset threshold, d. When F ≥ threshold d, it is considered to have passed Level 3 verification, and the total score becomes a + c + e (e being the Level 3 verification score).

[0041] S5-3: Credibility Calculation: Set total score thresholds f and g. When the total score ≥ threshold f, it is marked as "High Credibility"; when threshold g ≤ total score < threshold f, it is marked as "Medium Credibility"; when the total score < threshold g, it is marked as "Low Credibility", where g... <f。

[0042] Example 2: An open-source information collection and integration system based on geographic grid coding, which can implement the open-source information collection and integration method based on geographic grid coding described in Embodiment 1, includes: Information acquisition module: Acquires the geographical region to be divided and the information to be determined; Grid coding module: After dividing the geographic region to be divided into a geospatial grid hierarchy, each grid unit is geogrid coded. The graph generation module: After storing all information elements within the geographic grid based on the geographic grid encoding, it extracts entity relationships to generate a triplet knowledge graph; Data mapping module: performs multi-source heterogeneous data cleaning and disambiguation, as well as event time mapping, on the geographic grid after generating the triplet knowledge graph; Judgment and Integration Module: Based on the geographic grid after the event time mapping has been completed, the module performs correlation and credibility determination on the information to be judged, and completes open source information collection and integration based on the judgment results.

[0043] Example 3: This invention also provides an open-source information collection and integration device based on geographic grid coding, which can realize the open-source information collection and integration method based on geographic grid coding described in Embodiment 1, including a processor and a storage medium; The storage medium is used to store instructions; The processor is configured to operate according to the instructions to perform the steps of the following method: Obtain the geographical region to be divided and the information to be determined; After dividing the geographic region into a geospatial grid hierarchy, each grid unit is geocoded. After storing all information elements within the geographic grid based on the geographic grid encoding, entity relationships are extracted to generate a triplet knowledge graph. Multi-source heterogeneous data cleaning and disambiguation, as well as event time mapping, are performed on the geographic grid after the triplet knowledge graph is generated. Based on the geographic grid after the event time mapping has been completed, the correlation and credibility of the information to be judged are determined, and the open source information is collected and integrated based on the judgment results.

[0044] Example 4: This invention also provides a computer-readable storage medium that implements the open-source information collection and integration method based on geographic grid coding described in Embodiment 1. The medium stores a computer program that, when executed by a processor, performs the steps of the following method: Obtain the geographical region to be divided and the information to be determined; After dividing the geographic region into a geospatial grid hierarchy, each grid unit is geocoded. After storing all information elements within the geographic grid based on the geographic grid encoding, entity relationships are extracted to generate a triplet knowledge graph. Multi-source heterogeneous data cleaning and disambiguation, as well as event time mapping, are performed on the geographic grid after the triplet knowledge graph is generated. Based on the geographic grid after the event time mapping has been completed, the correlation and credibility of the information to be judged are determined, and the open source information is collected and integrated based on the judgment results.

[0045] As is known from common technical knowledge, this invention can be implemented through other embodiments that do not depart from its spirit or essential characteristics. Therefore, the disclosed embodiments described above are merely illustrative and not exhaustive. All modifications within the scope of this invention or its equivalents are included in this invention.

[0046] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0047] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0048] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0049] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0050] 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 it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. An open-source information collection and integration method based on geographic grid coding, characterized in that, include: Obtain the geographical region to be divided and the information to be determined; After dividing the geographic region into a geospatial grid hierarchy, each grid unit is geocoded. After storing all information elements within the geographic grid based on the geographic grid encoding, entity relationships are extracted to generate a triplet knowledge graph. Multi-source heterogeneous data cleaning and disambiguation, as well as event time mapping, are performed on the geographic grid after the triplet knowledge graph is generated. Based on the geographic grid after the event time mapping has been completed, the correlation and credibility of the information to be judged are determined, and the open source information is collected and integrated based on the judgment results.

2. The open-source information collection and integration method based on geographic grid coding according to claim 1, characterized in that, After dividing the geographic region into a geospatial grid hierarchy, each grid unit is geocoded, including: dividing the Earth's surface into multiple levels from high to low, where the longitude and latitude spans of the k-th level are: Dl k =360 / (2^k×m) Df k =180 / (2^k×n) In the formula: Δλ k For longitude span, Δφ k Where k is the latitudinal span, m and n are the number of grid levels, and m and n are constants related to the grid level. For each grid cell, it is encoded according to hierarchical information, longitude index, and latitude index, as shown in the expression: ID = k × 10^(2L) + i × 10^L + j i=floor((λ-λ0) / Δλ k ) j=floor((φ-φ0) / Δφ k ) In the formula: ID is the unique code of the grid, L is the maximum number of bits for indices i and j, i is the index in the longitude direction, j is the index in the latitude direction, λ is the longitude of the grid cell center, φ is the latitude of the grid cell center, λ0 is the starting longitude, φ0 is the starting latitude, and floor is the floor function.

3. The open-source information collection and integration method based on geographic grid coding according to claim 2, characterized in that, After storing all information elements within the geographic grid using geogrid encoding, entity relations are extracted to generate a triplet knowledge graph, including: Using the grid ID as the key, all information elements within the grid are stored, establishing a correspondence between the grid and the information elements. The information elements include domain names, IP addresses, social media accounts, and event reports. Entity relations are extracted from text / images using a BERT-based entity recognition and relation extraction model. The extracted results are then used to generate a triplet knowledge graph containing grid IDs, topic words, and confidence scores. The confidence score is calculated using the following formula: C = α × P + β × R Where C is the confidence level, P is the accuracy of entity relation extraction, R is the ratio of the frequency of the information element appearing in the grid to the total frequency, and α and β are weighting coefficients with α+β=1.

4. The open-source information collection and integration method based on geographic grid coding according to claim 3, characterized in that, The generated triplet knowledge graph mesh undergoes multi-source heterogeneous data cleaning and disambiguation, as well as event-time mapping, including: By integrating geocoding APIs from Gaode, Baidu, and Google, as well as a self-built postal code / POI knowledge graph, geocoding is performed on each API for the address information to be processed, resulting in multiple candidate coordinates. The similarity between each candidate coordinate and the corresponding address coordinates in the self-built knowledge graph is calculated, and the coordinate with the highest similarity is selected as the final coordinate of the address. A temporal convolutional network is used to predict the actual occurrence time of the event. The input of the temporal convolutional network is a feature vector containing the publication time and time descriptions in the event-related text, and the output is the time interval of the event. The predicted event time interval is mapped to the corresponding grid. The mapping rule is: if there is valid time coverage of the grid within the event occurrence time interval, then the event is mapped to that grid. The similarity calculation formula is as follows: S=1-d / D max In the formula: S is the similarity, d is the distance between the two coordinates, and D is the similarity score. max This is the maximum distance threshold that is set.

5. The open-source information collection and integration method based on geographic grid coding according to claim 4, characterized in that, Based on the geographic grid with completed event time mapping, the correlation and credibility of the information to be judged are determined, including: Calculate the temporal and geographical similarity of the information to be judged based on the geographic grid after the event time mapping has been completed; The spatiotemporal similarity of the information to be judged is calculated based on the temporal similarity and geographical similarity of the information to be judged. Calculate the degree of information association between the information to be judged based on the spatiotemporal similarity of the information to be judged; The degree of information correlation of the information to be judged is compared with a set threshold. If the degree of information correlation is greater than the set threshold, it is judged that the information has a correlation; otherwise, it is judged that there is no correlation. The credibility of the information to be judged is determined by calculating the total credibility score of the information to be judged through a three-level verification mechanism.

6. The open-source information collection and integration method based on geographic grid coding according to claim 5, characterized in that, The formula for calculating the degree of information correlation is: R= ×Sst+(1- )×Ss Ss=V A V B / (||V A ||×||V B ||) Sst = γ × St + (1 - γ) × Sg St=1-|t A -t B | / T max Sg=1-d AB / G max In the formula: R represents the degree of information correlation. Here, Sst represents the weighting coefficients for spatiotemporal similarity, Ss represents semantic similarity, and V represents the weighting coefficients for spatiotemporal similarity. A Let V be the semantic feature vector of information A. B Let γ be the semantic feature vector of information B, γ be the weight coefficient of temporal similarity, St be the temporal similarity, and t be the value of t. A Let t be the timestamp of information A. B Let T be the timestamp of information B. max The maximum time difference threshold is set, Sg is the geographical similarity, and d is the maximum time difference threshold. AB G represents the distance between the geographic coordinates of information A and information B. max This is the maximum geographic distance threshold that is set.

7. The open-source information collection and integration method based on geographic grid coding according to claim 6, characterized in that, The three-level verification mechanism includes primary verification, secondary verification, and tertiary verification, wherein: The first-level verification includes checking whether the information appears on at least two different types of open source platforms. If it appears, it is recorded as passing the first-level verification and a first-level verification score is assigned. The secondary verification includes, based on the primary verification, analyzing whether the key elements in the information are consistent, and calculating the consistency score C of the key elements. key When C key When the value is greater than or equal to the preset threshold b, it is considered as passing the second-level verification, and a second-level verification score is assigned and the first-level verification score is added. The three-level verification includes, based on passing the two-level verification, combining historical data and domain knowledge, judging whether the information conforms to logic and common sense, calculating the conformity degree F through logical reasoning and knowledge matching, and recording the passing of the three-level verification when F ≥ a preset threshold d, assigning a three-level verification score and accumulating the scores of the first and second-level verification as the total score. When the total score is greater than or equal to the preset total score threshold f, it is marked as "high confidence"; when the preset total score threshold g is less than the preset total score threshold f, it is marked as "medium confidence"; when the total score is less than the preset total score threshold g, it is marked as "low confidence", where the preset total score threshold g is less than the preset total score threshold f.

8. An open-source information collection and integration system based on geographic grid coding, characterized in that, include: Information acquisition module: Acquires the geographical region to be divided and the information to be determined; Grid coding module: After dividing the geographic region to be divided into a geospatial grid hierarchy, each grid unit is geogrid coded. The graph generation module: After storing all information elements within the geographic grid based on the geographic grid encoding, it extracts entity relationships to generate a triplet knowledge graph; Data mapping module: performs multi-source heterogeneous data cleaning and disambiguation, as well as event time mapping, on the geographic grid after generating the triplet knowledge graph; Judgment and Integration Module: Based on the geographic grid after the event time mapping has been completed, the module performs correlation and credibility determination on the information to be judged, and completes open source information collection and integration based on the judgment results.

9. An open-source information collection and integration device based on geographic grid coding, characterized in that, Including processor and storage media; The storage medium is used to store instructions; The processor is configured to operate according to the instructions to perform the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the method according to any one of claims 1 to 7.

Images