Geographic event knowledge graph alignment method and device, equipment and medium

Abstract

The invention belongs to the technical field of geographic information, and provides a geographic event knowledge graph alignment method and device, equipment and a medium, and the method comprises the steps: constructing a spatial neighborhood window of multiple scales, and obtaining multi-source heterogeneous data of the spatial neighborhood window; dividing the spatial neighborhood window into spatial semantic units; aggregating and classifying the multi-source attribute data of the spatial semantic unit to obtain an attribute vector and a structured tag of the spatial semantic unit, and further constructing a multi-scale scene semantic structure chart; encoding the multi-scale scene semantic structure chart, and performing multi-layer aggregation and updating on the multi-scale scene semantic structure chart through an encoding result to obtain scene semantic feature representation; and performing similarity clustering on the scene semantic feature representation to obtain a knowledge graph alignment result. The method has the beneficial effect of high geographic disaster identification accuracy.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of geographic information, and in particular to a geographic event knowledge graph alignment method, device, equipment and medium. BACKGROUND

[0002] In the prior art, knowledge graph entity alignment (KG entity alignment) has become one of the key technologies for solving the problem of heterogeneous semantic fusion, and its main goal is to identify two or more graph entities that are semantically equivalent or highly related, so as to realize the integration and reasoning enhancement of cross-source data. With the continuous growth of multi-source heterogeneous data, the importance of entity alignment is increasingly prominent, especially in the geographic space and disaster scenarios, accurate fusion of cross-source information is the basis for realizing complex relationship modeling and knowledge discovery.

[0003] However, the existing method still faces significant difficulties when applying knowledge graph alignment technology to the geographic event scenario. Many knowledge graph embedding models, such as SE-KGE, directly encode the geographic coordinates or boundary information of entities into the embedding space to improve spatial reasoning capabilities. However, such methods only rely on the geometric relationship between point coordinates and do not depict the complex spatial environmental characteristics of landslides, such as slope, land use distribution, fault structure, and vegetation coverage. Therefore, although coordinates can express location, they cannot reflect the semantic background information of event occurrence.

[0004] Some current research attempts to align entities with spatio-temporal features in the knowledge graph, and the existing technology has the following disadvantages: lack of structural modeling ability in geographic semantic scenarios; unable to identify spatially dispersed but environmentally similar landslide event pairs; lack of transition ability from spatial alignment to cognitive modeling. SUMMARY

[0005] To at least solve one of the technical problems existing in the prior art, the present application provides a geographic event knowledge graph alignment method, device, equipment and medium, which improves the accuracy of geographic disaster identification.

[0006] One aspect of the present application provides a geographic event knowledge graph alignment method, comprising: According to the target geographic event, a plurality of scale spatial neighborhood windows are constructed, and multi-source heterogeneous data of the spatial neighborhood windows is obtained; According to the digital elevation data of the spatial neighborhood window, the spatial neighborhood window is divided into spatial semantic units; The multi-source attribute data of the spatial semantic unit is aggregated and classified to obtain the attribute vector and structured label of the spatial semantic unit; The center point of the target geographic event is taken as an anchor node, and a multi-scale scene semantic structure graph is determined according to the anchor node, spatial semantic units adjacent to the anchor node and attribute vectors of the spatial semantic units; The multi-scale scene semantic structure graph is encoded, and multi-layer aggregation and updating of the multi-scale scene semantic structure graph are performed through an encoding result to obtain a scene semantic feature representation; The scene semantic feature representation is subjected to similarity clustering to obtain a knowledge graph alignment result.

[0007] According to the geographic event knowledge graph alignment method, a plurality of scale spatial neighborhood windows are constructed according to the target geographic event, and multi-source heterogeneous data of the spatial neighborhood windows are obtained, including: The target geographic event is taken as a center point, and a plurality of spatial neighborhood windows of different scales are constructed; Multi-source heterogeneous data of each scale of the spatial neighborhood windows are collected and integrated, wherein the multi-source heterogeneous data includes monitoring data and event factor data.

[0008] According to the geographic event knowledge graph alignment method, digital elevation data of the spatial neighborhood window is used to divide the spatial neighborhood window into spatial semantic units, including: The digital elevation data is used to divide the spatial neighborhood window into a plurality of spatial semantic units by ARCGIS, and boundaries between adjacent spatial semantic units are demarcated by a watershed or a topographic feature.

[0009] According to the geographic event knowledge graph alignment method, multi-source attribute data of the spatial semantic unit are aggregated and classified to obtain an attribute vector and a structured label of the spatial semantic unit, including: The multi-source attribute data of each spatial semantic unit are aggregated to obtain an attribute vector of the spatial semantic unit; The attribute vector of the spatial semantic unit is used to perform semantic classification by using event professional knowledge and a preset threshold rule to obtain a structured label of the spatial semantic unit.

[0010] According to the geographic event knowledge graph alignment method, the center point of the target geographic event is taken as an anchor node, and a multi-scale scene semantic structure graph is determined according to the anchor node, spatial semantic units adjacent to the anchor node and attribute vectors of the spatial semantic units, including: The spatial semantic unit in which the center point of the target geographic event is located is taken as an anchor node; Spatial semantic units adjacent to the anchor node are taken as unit nodes, and a topological edge between the anchor node and the unit node is determined, wherein the type of the topological edge includes grid adjacency, slope aspect continuity and tectonic unit connectivity; Determine a multi-scale scene semantic structure graph according to attribute vectors of anchor nodes, unit nodes, topology edges and spatial semantic units For: Among them, is a node set composed of anchor nodes and unit nodes, is a set of topology edges, is a set of attribute vectors of spatial semantic units.

[0011] According to the geographical event knowledge graph alignment method, the multi-scale scene semantic structure graph is encoded, and the multi-scale scene semantic structure graph is multi-layer aggregated and updated through the encoding result to obtain a scene semantic feature representation, including: According to the attribute type of the anchor node and the unit node in the multi-scale scene semantic structure graph, the encoding processing is performed to obtain a node initial representation, wherein the encoding includes: The numerical attribute of the anchor node and the unit node is processed by normalization to obtain a numerical vector; The structured label of the anchor node and the unit node is converted into a semantic embedding vector by using a bert pre-training model; The numerical vector and the semantic embedding vector are spliced to obtain the node initial representation of the node Among them is a numerical vector, is a semantic embedding vector; According to the node initial representation, the multi-scale scene semantic structure graph is multi-layer aggregated and updated to obtain a node final representation Among them, the formula for multi-layer aggregation and updating of the multi-scale scene semantic structure graph is: Among them is a node vector, is a weight for controlling the characteristics of the node itself and the neighbor characteristics, represents the initial feature vector of the node , and represents the neighbor node set of the node .

[0012] According to the geographical event knowledge graph alignment method, the scene semantic feature representation is similarity clustered to obtain a knowledge graph alignment result, including: The node final representation is weighted and averaged to obtain a graph-level vector is: Among them is a weight, is a weight.​​ is the total number of node vectors; The multi-scale scene semantic structure graph is identified by a distance threshold and a density clustering method to obtain events or event clusters with similar spatial semantics, and a knowledge graph alignment result is obtained. Event nodes in the same type of scene are obtained from the events or event clusters with similar spatial semantics, and an aligned event node list is obtained.

[0013] Another aspect of the present application provides a geographic event knowledge graph alignment device, comprising: A first module is configured to construct a plurality of scale spatial neighborhood windows according to a target geographic event, and obtain multi-source heterogeneous data of the spatial neighborhood windows. A second module is configured to divide the spatial neighborhood windows into spatial semantic units according to digital elevation data of the spatial neighborhood windows. A third module is configured to aggregate and classify multi-source attribute data of the spatial semantic units to obtain attribute vectors and structured labels of the spatial semantic units. A fourth module is configured to take a center point of the target geographic event as an anchor node, and determine a multi-scale scene semantic structure graph according to the anchor node, spatial semantic units adjacent to the anchor node, and attribute vectors of the spatial semantic units. A fifth module is configured to encode the multi-scale scene semantic structure graph, and perform multi-layer aggregation and update on the multi-scale scene semantic structure graph through the encoding result to obtain a scene semantic feature representation. A sixth module is configured to perform similarity clustering on the scene semantic feature representation to obtain a knowledge graph alignment result.

[0014] Another aspect of the present application provides an electronic device comprising a processor and a memory. The memory is configured to store a program. The processor executes the program to implement the method described above.

[0015] The present application also discloses a computer program product or computer program comprising computer instructions stored in a computer readable storage medium. The processor of the computer device can read the computer instructions from the computer readable storage medium, and the processor executes the computer instructions to make the computer device execute the method described above.

[0016] The application has the beneficial effects that: by constructing a graph structure based on a spatial semantic unit, introducing adjacency, constructing topological relations such as connectivity, the spatial organization structure and geographic semantics of the environment where the event is located are effectively expressed; by setting a plurality of neighborhood windows of different spatial scales, the environment structure information is extracted at different spatial perception levels, the problem of incomplete or oversimplified semantic expression under a fixed scale is solved, and the ability to depict the heterogeneity of the geographic event environment is improved; by embedding the graph structure to obtain a scene structure vector, matching and clustering are performed based on structural similarity, the spatial proximity assumption is broken, and the cross-regional and non-local event alignment requirements are supported; the spatial semantic structure graph construction and alignment framework are not dependent on specific event types, can be extended to scene recognition and merging tasks of various geographic phenomena, and has high application and promotion potential. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 is a geographic event knowledge graph alignment process schematic diagram of the embodiment of the application; Figure 2 is a spatial neighborhood window construction process schematic diagram of the embodiment of the application; Figure 3 is a multi-source attribute data aggregation and classification process schematic diagram of the embodiment of the application; Figure 4 is a multi-scale scene semantic structure graph construction process schematic diagram of the embodiment of the application; Figure 5 is a scene semantic feature generation process schematic diagram of the embodiment of the application; Figure 6 is a scene semantic feature generation process schematic diagram of the embodiment of the application; Figure 7 is a knowledge graph alignment process schematic diagram of the embodiment of the application; Figure 8 is a knowledge graph alignment method flowchart of hidden danger events and landslide events of the embodiment of the application; Figure 9 is a geographic event knowledge graph alignment device schematic diagram of the embodiment of the application. DETAILED DESCRIPTION

[0018] Embodiments of the present application are described below in detail with reference to the accompanying drawings, examples of which are shown in the drawings, wherein the same or similar numerals represent the same or similar elements or elements having the same or similar functions throughout the drawings. In the following description, the suffix "module", "part" or "unit" used for an element is merely intended for facilitating explanation of the present application, and by itself does not have a specific meaning. Therefore, "module", "part" or "unit" can be mixedly used. "First", "second", and so on are used only to distinguish technical features for the purpose of explanation, and cannot be understood to indicate or imply relative importance or implicitly indicate the number of indicated technical features or the order of the indicated technical features. In the following description, the consecutive numerals of the method steps are for the convenience of review and understanding, and adjusting the implementation order between the steps does not affect the technical effects achieved by the technical solution of the present application in combination with the overall technical solution of the present application and the logical relationship between the steps. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present application, and cannot be understood as a limitation on the present application.

[0019] Reference Figure 1 , Figure 1 is a geographic event knowledge graph alignment method flowchart of an embodiment of the present application, which includes but is not limited to steps S100-S600: S100, according to a target geographic event, constructing a plurality of scale spatial neighborhood windows and acquiring multi-source heterogeneous data of the spatial neighborhood windows.

[0020] In some embodiments, the spatial neighborhood window construction flowchart shown in Figure 2 includes but is not limited to steps S110-S120: S110, constructing a plurality of spatial neighborhood windows of different scales with the target geographic event as the center point; S120, collecting and integrating multi-source heterogeneous data of each scale of the spatial neighborhood windows, wherein the multi-source heterogeneous data includes monitoring data and event factor data.

[0021] In some embodiments, taking a landslide event as an example, a plurality of scale spatial neighborhood windows are constructed respectively with each landslide disaster point or landslide danger point as the center, for example, concentric circular areas with radii of 60m, 180m, 600m and 1km are taken as spatial neighborhood windows, which respectively correspond to the spatial environmental influence range of small, medium, large and giant landslides, for acquiring the environmental context of the event under different scales.

[0022] In some embodiments, within each spatial neighborhood window, multi-source heterogeneous geographic data is collected and integrated, wherein the monitoring data includes coverage rainfall and displacement, and the event factor data includes landform partition, earthquake, DEM (digital elevation data) and land use type, etc.

[0023] It should be noted that the event factor data is used to represent data causing a geographical event (such as a landslide, a flash flood).

[0024] S200, according to the digital elevation data of the spatial neighborhood window, the spatial neighborhood window is divided into spatial semantic units.

[0025] In some embodiments, according to the digital elevation data (DEM), the spatial neighborhood window is divided into a plurality of spatial semantic units by ARCGIS, and the boundaries between adjacent spatial semantic units are demarcated by a watershed or a topographic feature.

[0026] For example, taking a landslide event as an example, the DEM data in the neighborhood window is divided into a plurality of natural slope units covering the neighborhood by the slope unit segmentation of the topographic analysis tool of ARCGIS, wherein the natural slope unit is a spatial semantic unit representing the topography.

[0027] S300, the multi-source attribute data of the spatial semantic unit is aggregated and classified to obtain the attribute vector and the structured label of the spatial semantic unit.

[0028] In some embodiments, referring to Figure 3 The aggregation and classification process diagram of the multi-source attribute data is shown, which includes but is not limited to steps S310-S320: S310, the multi-source attribute data of each spatial semantic unit is aggregated to obtain the attribute vector of the spatial semantic unit; S320, according to the attribute vector of the spatial semantic unit, the semantic classification is performed by using event professional knowledge and a preset threshold rule to obtain the structured label of the spatial semantic unit.

[0029] In some embodiments, the multi-source attribute data in each slope unit (spatial semantic unit) is aggregated, including average slope, average slope direction, NDVI average, land use type proportion, fault density, geological unit type, etc., and finally the attribute vector of the slope unit (spatial semantic unit) is formed, wherein the attribute vector of the slope unit (spatial semantic unit) is used as the node feature of the subsequent node (spatial semantic unit).

[0030] In some embodiments, according to the slope unit attribute, the semantic classification is performed by using the landslide disaster professional knowledge and the existing threshold rule to obtain the structured label of the slope unit (spatial semantic unit), such as steep slope forest land, gentle slope farmland, fault edge area and collapse valley, etc. The structured label enhances the semantic expression ability of the subsequent node.

[0031] S400, taking the center point of the target geographic event as an anchor node, determining a multi-scale scene semantic structure graph according to the anchor node, spatial semantic units adjacent to the anchor node, and attribute vectors of the spatial semantic units.

[0032] In some embodiments, referring to Figure 4 a multi-scale scene semantic structure graph construction flowchart shown, which includes but is not limited to steps S410-S430: S410, taking the spatial semantic unit where the center point of the target geographic event is located as an anchor node; S420, taking the spatial semantic units adjacent to the anchor node as unit nodes, and determining the topological edges between the anchor node and the unit nodes, wherein the types of the topological edges include grid adjacency, slope aspect continuity, and structural unit connectivity; S430, determining a multi-scale scene semantic structure graph according to the anchor node, the unit nodes, the topological edges, and the attribute vectors of the spatial semantic units.

[0033] In some embodiments, the multi-scale scene semantic structure graph is: ; wherein, the node set is composed of the anchor node and the unit nodes, the topological edge set is composed of the topological edges, and the attribute vector set is composed of the attribute vectors of the spatial semantic units.

[0034] It should be noted that the attribute vector of the spatial semantic unit represents the node features in the node set, and represents the clustered multi-source attribute data shown in the above embodiments.

[0035] S500, encoding the multi-scale scene semantic structure graph, and performing multi-layer aggregation and updating on the multi-scale scene semantic structure graph through the encoding result to obtain a scene semantic feature representation.

[0036] In some embodiments, referring to Figure 5 a scene semantic feature generation flowchart shown, which includes but is not limited to steps S510-S520: S510, performing encoding processing according to the attribute types of the anchor nodes and the unit nodes in the multi-scale scene semantic structure graph to obtain node initial representations; In some embodiments, wherein the encoding includes: obtaining a numerical vector by normalizing the numerical attributes of the anchor nodes and the unit nodes; converting the structured labels of the anchor nodes and the unit nodes into semantic embedding vectors using a bert pre-training model; and concatenating the numerical vector and the semantic embedding vector to obtain the node initial representation of the node wherein, the numerical vector is This is a semantic embedding vector.

[0037] S520: Based on the initial node representation, multi-level aggregation and updating of the multi-scale scene semantic structure graph are performed to obtain the final node representation; In some embodiments, a node is ultimately represented as The formula for multi-level aggregation and updating of the semantic structure graph of a multi-scale scene is as follows: ; in, For node vectors, To control the weights of a node's own features and those of its neighbors; Representative node The initial feature vector, the numerical attributes of the nodes, and the semantic label information; Represents a node The set of neighboring nodes, that is, the set of nodes with which the node is located. A set of nodes connected by topological edges in a multi-scale scene semantic structure graph.

[0038] In some embodiments, the scene semantic feature vector embedding is performed by the above formula, the scene semantic structure graph is aggregated and updated in multiple layers, node attributes and topology information are fused, and a fixed-dimensional embedding vector of the event graph as a whole is generated, which accurately represents the geographic scene semantic features of physical events.

[0039] S600 performs similarity clustering on the semantic feature representation of the scene to obtain the knowledge graph alignment result.

[0040] In some embodiments, reference Figure 6 The diagram showing the image recognition and alignment process includes, but is not limited to, steps S610 to S630: S610, perform weighted average aggregation on the final node representations to obtain graph-level vectors; S620 uses a distance threshold and density clustering method to identify multi-scale scene semantic structure graphs, obtain spatially semantically similar events or event clusters, and obtain knowledge graph alignment results; S630: Obtain event nodes of the same type of scenario from events or event clusters with similar spatial semantics, and then obtain an aligned list of event nodes.

[0041] Among them, graph-level vectors for: ; in, As weight, This represents the total number of node vectors.

[0042] It should be noted that the graph-level vector implemented in the present application is used to integrate the scattered node features in the graph into a unified vector representation.

[0043] In some embodiments, the similarity of the embedding vectors of all landslide disaster event graphs and landslide hazard event graphs is calculated, and spatially and semantically similar event pairs or event clusters are identified based on a distance threshold and a density-based clustering (DBSCAN) method, so as to realize the alignment of knowledge graphs between different data sources and different types of landslide events.

[0044] In some embodiments, according to the event nodes corresponding to the vectors in the clustering cluster, event nodes belonging to the same type of scene are extracted, and an aligned event node ID list is output.

[0045] In some embodiments, reference is made to Figure 7 and Figure 8 , wherein Figure 7 is a schematic diagram of a geographic event knowledge graph alignment model, and Figure 8 is a flowchart of a hidden danger event and landslide event knowledge graph alignment method, in combination with Figure 7 and Figure 8 , the process is as follows: (1) Obtain hidden danger event and landslide event point data.

[0046] (2) Reference the event neighborhood division into concentric circle regions of 60m, 180m, 600m and 1km according to the definition of landslide event scale.

[0047] (3) Use DEM data to construct slope units and aggregate multi-factor attributes as spatial semantic entities (spatial semantic units) carrying events, i.e., obtain landslide spatial attribute data and event spatio-temporal position.

[0048] (4) According to rainfall, displacement, etc. monitoring data, landform partition, earthquake, DEM, land use type, etc. (event) factor data, use DEM data to construct slope units and aggregate multi-factor attributes as spatial semantic entities carrying events; then use spatial overlay analysis to construct the basic spatial topological relationship between entities, model hydrological relationship based on DEM, and model other semantic relationships based on priori mechanism knowledge.

[0049] (5) Construct event subgraphs in the manner of the above embodiments. It should be noted that the event subgraphs are multi-scale scene semantic structure graphs constructed at different scales.

[0050] (6) Perform entity and event subgraph embedding calculation, vector clustering and similar scene event subgraph identification on each subgraph, and then obtain the aligned hidden danger-landslide event graph.

[0051] In some embodiments, the output result part example is shown in Table 1.

[0052] Table 1 output result part example table

[0053] The C-001 of the event cluster ID and the event cluster ID-002 in Table 1 have high similarity in knowledge graph alignment under different event and scene entity types, which proves that the technical scheme of the embodiment of the present application achieves a good mining effect of geographic events.

[0054] Figure 9 is a schematic diagram of a geographic event knowledge graph alignment device according to an embodiment of the present application. The device includes a first module 910, a second module 920, a third module 930, a fourth module 940, a fifth module 950, and a sixth module 960.

[0055] The first module is configured to construct a plurality of scale spatial neighborhood windows according to a target geographic event, and obtain multi-source heterogeneous data of the spatial neighborhood windows; the second module is configured to divide the spatial neighborhood windows into spatial semantic units according to digital elevation data of the spatial neighborhood windows; the third module is configured to aggregate and classify multi-source attribute data of the spatial semantic units to obtain attribute vectors and structured labels of the spatial semantic units; the fourth module is configured to take a center point of the target geographic event as an anchor node, and determine a multi-scale scene semantic structure graph according to the anchor node, spatial semantic units adjacent to the anchor node, and attribute vectors of the spatial semantic units; the fifth module is configured to encode the multi-scale scene semantic structure graph, and perform multi-layer aggregation and update on the multi-scale scene semantic structure graph through an encoding result to obtain scene semantic feature representation; and the sixth module is configured to perform similarity clustering on the scene semantic feature representation to obtain a knowledge graph alignment result.

[0056] Exemplarily, under cooperation of the first module, the second module, the third module, the fourth module, the fifth module and the sixth module in the device, the embodiment device can implement any one of the preceding geographic event knowledge graph alignment methods, that is, according to a target geographic event, a plurality of scale spatial neighborhood windows are constructed, and multi-source heterogeneous data of the spatial neighborhood windows are obtained; according to digital elevation data of the spatial neighborhood windows, the spatial neighborhood windows are divided into spatial semantic units; multi-source attribute data of the spatial semantic units are aggregated and classified to obtain attribute vectors and structured labels of the spatial semantic units; a center point of the target geographic event is taken as an anchor node, and according to the anchor node, spatial semantic units adjacent to the anchor node and the attribute vectors of the spatial semantic units, a multi-scale scene semantic structure graph is determined; the multi-scale scene semantic structure graph is encoded, and the multi-scale scene semantic structure graph is multi-layer aggregated and updated through an encoding result to obtain a scene semantic feature representation; the scene semantic feature representation is subjected to similarity clustering to obtain a knowledge graph alignment result. The present application has the beneficial effects that: by constructing a graph structure based on spatial semantic units, introducing adjacency, constructing topological relations such as connectivity, the spatial organization structure of the environment where the event is located and the geographic semantics thereof are effectively expressed; by setting a plurality of neighborhood windows of different spatial scales, environmental structure information is extracted at different spatial perception levels, the problem of incomplete or oversimplified semantic expression under a fixed scale is solved, and the ability to depict geographic event environment heterogeneity is improved; by obtaining a scene structure vector through graph structure embedding, matching and clustering based on structural similarity, the spatial proximity assumption is broken, and cross-regional and non-local event alignment requirements are supported; the spatial semantic structure graph construction and alignment framework is not dependent on a specific event type, can be extended to scene recognition and merging tasks of various geographic phenomena, and has high application and promotion potential.

[0057] The embodiment of the present application further provides an electronic device, which comprises a processor and a memory; The memory stores a program; The processor executes the program to perform the aforementioned geographic event knowledge graph alignment method; the electronic device has the function of carrying and running the software system for geographic event knowledge graph alignment provided by the embodiment of the present application, for example, a personal computer, a mini computer, a mainframe, a workstation, a network or a distributed computing environment, a separate or integrated computer platform, or communication with a charged particle tool or other imaging device, and the like.

[0058] The embodiment of the present application further provides a computer readable storage medium, which stores a program, and the program is executed by a processor to implement the geographic event knowledge graph alignment method as described above.

[0059] In some alternative embodiments, the functions / operations described in the block diagrams can not occur in the order presented in the operational illustrations. For example, two blocks shown in succession can in fact be executed substantially concurrently or the blocks can sometimes be executed in the reverse order, depending upon the functionality / operations involved. Also, although the embodiments presented in the flow diagrams have been described in connection with the embodiments presented in the block diagrams, it should be understood that many other embodiments can be implemented in accordance with the present application. For example, an object, which is employed in a software implementation, can be implemented in hardware, or with both hardware and software having one or more functions and operations of the object combined in a common device. The described operations and their associated logic flow are meant to be examples. Depending on the implementation, acts sequential to some other acts can be performed, or some acts can be performed concurrently, or even at a later date, especially where it is understood that the order of acts can vary and further acts can be added. Accordingly, not all of the acts are required, but include the functions and operations essential for the practice of the present application.

[0060] The embodiments of the present application further disclose a computer program product or computer program, which comprises computer instructions stored in a computer readable storage medium. A processor of a computer device can read the computer instructions from the computer readable storage medium, and the processor executes the computer instructions, so that the computer device executes the aforementioned method for aligning a geographical event knowledge graph.

[0061] Further, while the present application has been described in the context of functional modules, it is to be understood that one or more of the described functions and / or features can be integrated in a single physical device and / or software module, or one or more functions and / or features can be implemented in separate physical devices or software modules. It will also be appreciated that detailed discussion of the actual implementation of each module is not necessary to an understanding of the application. Rather, the actual implementation of the modules, in combination with their attributes, functions, and internal relationships, as disclosed in the embodiments of the present application, is within the skill of the engineer in charge of such implementations. Accordingly, the present application is not limited to the embodiments disclosed herein, but is intended to cover any and all adaptations of these embodiments. It is also to be understood that the disclosed specific concepts are merely illustrative and are not intended to limit the scope of the present application as defined by the appended claims and equivalents thereof.

[0062] If the functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application or the parts of the present application that essentially contribute to the prior art or the parts of the technical solutions of the present application can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0063] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a list of executable instructions for implementing logic functions, which can be embodied in any computer readable medium for use by or in connection with an instruction execution system, apparatus or device, such as a computer-based system, a system including a processor or other system that can fetch instructions from an instruction execution system, apparatus or device and execute the instructions, or in conjunction with these instruction execution systems, apparatus or devices. For the purpose of this specification, "computer readable medium" can be any device that can include, store, communicate, propagate or transport programs for use by or in connection with an instruction execution system, apparatus or device, or in conjunction with these instruction execution systems, apparatus or devices.

[0064] More specific examples (non-exhaustive list) of the computer readable medium include the following: an electrical connection having one or more wires (electrical devices), a portable computer diskette (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CD ROM). In addition, the computer readable medium can even be paper or other suitable medium on which the program can be printed, because the program can be obtained electronically, for example, by optical scanning of the paper or other medium, followed by editing, interpreting or otherwise processing, if necessary, in other suitable ways, to be electronically obtained and then stored in the computer memory.

[0065] It should be understood that aspects of the application can be implemented in hardware, software, firmware or a combination thereof. In the above embodiments, various steps or methods can be implemented in software or firmware which is stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, and in another embodiment, any of the following technologies, known in the art, can be used: a combination of discrete logic circuits having logic gates for implementing logic functions upon an application of data signals, application specific integrated circuits having logic gates, field programmable gate arrays (FPGA), or others.

[0066] In the description of the present specification, the description of the terms "one embodiment", "some embodiments", "an example", "a specific example", or "some examples" and the like means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In the present specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Also, the specific features, structures, materials or characteristics described can be combined in any appropriate manner in one or more embodiments or examples.

[0067] Although the embodiments of the present application have been shown and described, it will be understood by those skilled in the art that various changes, modifications, alternatives and variations can be made thereto without departing from the principles and spirit of the present application, the scope of which is defined by the appended claims and their equivalents.

[0068] The above is a specific description of the preferred embodiments of the present application, but the present application is not limited to the described embodiments, and those skilled in the art can make various equivalent modifications or replacements without departing from the spirit of the present application, and these equivalent modifications or replacements are included in the scope defined by the claims of the present application.

Claims

1. A method for aligning geographical event knowledge graphs, characterized in that, The method comprises the following steps: According to the target geographical event, a plurality of scale spatial neighborhood windows are constructed, and multi-source heterogeneous data of the spatial neighborhood windows are obtained; According to the digital elevation data of the spatial neighborhood window, the spatial neighborhood window is divided into spatial semantic units; The multi-source attribute data of the spatial semantic units are aggregated and classified to obtain the attribute vector and the structured label of the spatial semantic units; The center point of the target geographical event is taken as an anchor node, and a multi-scale scene semantic structure graph is determined according to the anchor node, the spatial semantic units adjacent to the anchor node and the attribute vector of the spatial semantic units; The multi-scale scene semantic structure graph is encoded, and the multi-scale scene semantic structure graph is aggregated and updated in multiple layers through the encoding result to obtain a scene semantic feature representation; The scene semantic feature representation is subjected to similarity clustering to obtain a knowledge graph alignment result. 2.The method of aligning geographical event knowledge graphs according to claim 1, characterized in that, The method comprises the following steps: A plurality of spatial neighborhood windows of different scales are constructed with the target geographical event as a center point; Multi-source heterogeneous data of each scale of spatial neighborhood window are collected and integrated, wherein the multi-source heterogeneous data comprises monitoring data and event factor data. 3.The method of aligning geographical event knowledge graphs according to claim 1, characterized in that, The spatial neighborhood window is divided into a plurality of spatial semantic units according to the digital elevation data, and the boundaries between adjacent spatial semantic units are demarcated by a watershed or a topographic feature. The multi-source attribute data of the spatial semantic units are aggregated and classified to obtain the attribute vector and the structured label of the spatial semantic units, which comprises the following steps: 4.The method of aligning geographical event knowledge graphs according to claim 1, characterized in that, The multi-source attribute data of each spatial semantic unit are aggregated to obtain the attribute vector of the spatial semantic unit; The attribute vector of the spatial semantic unit is subjected to semantic classification by using event professional knowledge and a preset threshold rule to obtain the structured label of the spatial semantic unit. The center point of the target geographical event is taken as an anchor node, and a multi-scale scene semantic structure graph is determined according to the anchor node, the spatial semantic units adjacent to the anchor node and the attribute vector of the spatial semantic units, which comprises the following steps:

5. The method of claim 4, wherein, The spatial semantic unit in which the center point of the target geographical event is located is taken as an anchor node; The spatial semantic units adjacent to the anchor node are taken as unit nodes, and a topological edge between the anchor node and the unit node is determined, wherein the type of the topological edge comprises grid adjacency, slope aspect continuity and tectonic unit connectivity. The multi-scale scene semantic structure graph is encoded, and the multi-scale scene semantic structure graph is aggregated and updated in multiple layers through the encoding result to obtain a scene semantic feature representation, which comprises the following steps: Determining a multi-scale scene semantic graph according to attribute vectors of anchor nodes, unit nodes, topology edges, and spatial semantic units To: ； wherein, is a set of nodes consisting of anchor nodes and unit nodes, is a set of topological edges, is a set of attribute vectors of spatial semantic units.

6. The method of claim 5, wherein, According to the attribute type of the anchor node and the unit node in the multi-scale scene semantic structure graph, an encoding process is performed to obtain a node initial representation, wherein the encoding comprises: The numerical attribute of the anchor node and the unit node is subjected to normalization processing to obtain a numerical vector; The structured label of the anchor node and the unit node is converted into a semantic embedding vector by using a bert pre-training model. ​ concatenate the numerical vector and the semantic embedding vector to obtain a node initial representation of the node wherein is a numerical vector, is a semantic embedding vector; According to the node initial representation, multi-layer aggregation and updating are performed on the multi-scale scene semantic structure graph to obtain a node final representation wherein a formula for performing multi-layer aggregation and updating on the multi-scale scene semantic structure graph is: ； wherein, is a node vector, is a weight controlling the node's own feature and neighbor's feature, represents a node an initial feature vector of the node, denotes a set of neighbor nodes of the node .

7. The method of claim 6, wherein, The similarity clustering is performed on the scene semantic feature representation, and a knowledge graph alignment result is obtained, including: finalizing the node representation performing weighted average aggregation to obtain graph-level vectors is: ； wherein, is a weight, is the total number of node vectors; The multi-scale scene semantic structure graph is recognized by using a distance threshold and density clustering method, events or event clusters with similar spatial semantics are obtained, and the knowledge graph alignment result is obtained; Event nodes in the same type of scene are obtained from the events or event clusters with similar spatial semantics, and an aligned event node list is obtained.

8. A geographic event knowledge graph alignment apparatus, characterized in that, It includes: The first module is configured to construct a plurality of scale spatial neighborhood windows according to a target geographic event, and obtain multi-source heterogeneous data of the spatial neighborhood windows; The second module is configured to divide the spatial neighborhood windows into spatial semantic units according to digital elevation data of the spatial neighborhood windows; The third module is configured to aggregate and classify multi-source attribute data of the spatial semantic units to obtain attribute vectors and structured labels of the spatial semantic units; The fourth module is configured to take a center point of the target geographic event as an anchor node, and determine a multi-scale scene semantic structure graph according to the anchor node, spatial semantic units adjacent to the anchor node, and attribute vectors of the spatial semantic units; The fifth module is configured to encode the multi-scale scene semantic structure graph, and perform multi-layer aggregation and update on the multi-scale scene semantic structure graph through an encoding result to obtain a scene semantic feature representation; The sixth module is configured to perform similarity clustering on the scene semantic feature representation to obtain a knowledge graph alignment result.

9. An electronic device, comprising: It includes a processor and a memory; The memory is configured to store a program; The processor executes the program to implement the geographic event knowledge graph alignment method according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The storage medium stores a program, and the program is executed by the processor to implement the geographic event knowledge graph alignment method according to any one of claims 1-7.

Images