Space-time fusion coding method, data relation generation method and system

Abstract

The invention provides a space-time fusion coding method and a data relation generation method and system, and is applied to the technical field of physical world-oriented multi-source heterogeneous data organization management. Space information of a target data object is mapped to a geographic space subdivision grid coding system to generate a space code; the method comprises the following steps: standardizing time information according to preset time granularity (such as year, month, day and hour), mapping the time information to a layered time coding system to generate a time code, and fusing a space code, the time code and object identification information according to a preset rule to generate a unique space-time fusion code; and finally, the space-time fusion codes and attribute information are bound to form standardized data records, data objects of different sources and different formats are mapped to a unified space-time coding system, data islands are broken, and a data foundation is laid for constructing a unified space-time relation network.

Description

Technical Field

[0001] This application relates to the field of multi-source heterogeneous data management technology, specifically to a spatiotemporal fusion coding method, a data relationship generation method, and a system. Background Technology

[0002] With the widespread deployment of the Internet of Things (IoT), remote sensing, Building Information Modeling (BIM), urban management information systems, and enterprise business systems in governments, enterprises, and industries, massive amounts of multi-source, multi-modal, and spatiotemporal data are continuously generated and accumulated. These data are complex in form and diverse in source, and can be broadly categorized into the following three types: Structured data—such as time-series data from sensors / equipment, business tables in enterprise management systems (such as equipment ledgers, asset tables, operation and maintenance records, etc.), and databases of urban management systems—is recorded in a standardized manner with clearly defined fields. However, it often suffers from problems such as scattered source systems and inconsistent primary key systems, leading to difficulties in cross-system integration. Unstructured / semi-structured data—such as various documents (e.g., ownership certificates, contract texts), forms (Excel spreadsheets), images, videos, and CAD files—lacks a unified field representation, making it difficult to directly use for relational retrieval and analysis. This also includes geographic information data—such as BIM models, CAD plans, remote sensing images, and vector layers—which are information carriers with clear spatial references. While this type of data is rich in spatial semantics and geometric representation, in practical applications, issues such as inconsistencies in projection, resolution, and semantic labels create barriers to integration.

[0003] To support intelligent operation and maintenance, spatiotemporal behavior analysis, asset management, and decision support, the industry generally needs to fuse and correlate the aforementioned heterogeneous data sources across systems and modalities to obtain a "full attribute view of the same object / same spatial location at different points in time." Currently, commonly used data fusion methods and tools mainly include the following categories: The relational database / primary key mapping method achieves connections between tables of the same origin or that can be mapped by establishing a mapping table at the upper level or defining foreign key / primary key correspondences during data import. This method is effective for structured data, but lacks a direct mapping mechanism for a large number of unstructured and spatially weakly correlated files (such as images, Word documents, PDFs, XXLXS files, etc.), and it is difficult to express the uniqueness of the combination of time and spatial dimensions simultaneously.

[0004] Semantic / Ontology and Knowledge Graph Approach: This approach constructs domain ontology and entity-attribute-relationship (ER) models, and utilizes techniques such as entity alignment and text extraction to map multi-source data into a unified semantic layer (knowledge graph). While this method excels at expressing complex semantic relationships, it suffers from high construction and maintenance costs. Furthermore, it is prone to semantic inconsistencies and time-consuming manual intervention when aligning semantics across multiple languages ​​and formats. Additionally, knowledge graphs typically emphasize semantic abstraction, while their ability to express precise spatial geometric locations and time points / time periods, as well as their efficient indexing capabilities, are relatively weaker.

[0005] GIS fusion based on spatial coordinates / geometric indexes relies on latitude / longitude / projected coordinates, spatial buffers, and topological relationships (intersection, containment, etc.) to establish connections between geographic objects. This approach can effectively handle spatial relationships between geographic features, but when the data source is not spatially native (e.g., equipment logs, contract documents, semantic inconsistencies in 2D CAD), it is difficult to directly apply the data to the spatial index. Furthermore, inconsistencies in coordinate reference systems, precision, and timestamps across different data sources can lead to association errors and ambiguities.

[0006] Although the above methods have applications in their respective fields, the following problems still exist when facing the comprehensive application needs of cross-modal (structured / unstructured / geographic information) and strong spatiotemporal correlation: First, data location is disconnected from spatiotemporal scenarios: While traditional identifier systems (such as DOI in the field of scientific data and DSSN in the Internet of Data) can provide globally unique logical identifiers for data objects, their encoding itself lacks native spatial and temporal attributes and cannot be directly mapped to latitude and longitude coordinates and time points in the physical world, making it difficult to support spatiotemporally sensitive businesses (such as disaster emergency response and logistics tracking). In addition, traditional database encoding is mostly database storage primary key encoding, which does not have the characteristics of spatiotemporal big data primary keys, and cannot support efficient spatiotemporal big data analysis. Furthermore, the basic equipment lacks visualization, refined management, and clear hierarchy.

[0007] Second, the coexistence of multiple coding systems creates "data silos": different fields (such as scientific data, medical data, and data networks) use independent and incompatible identification and coding systems (such as DOI and DSSN), resulting in system fragmentation. Traditional coding often adopts a "type + sequence" structure, and the lack of uniformity in systems and coding methods between industries and departments leads to difficulties in information resource integration and significant obstacles to cross-departmental sharing and openness.

[0008] Third, the integration of non-spatiotemporal data is difficult: data spaces without embedded spatiotemporal coding (such as structured databases) are difficult to interact efficiently with spatiotemporal driven systems (such as BeiDou augmentation services). Moreover, core platforms are difficult to systematize, deployment and maintenance are difficult and costly, and there is a lack of complete methods to generate a unique spatiotemporal identifier for each object and to build relationships between attributes based on this identifier system.

[0009] Therefore, a new spatiotemporal fusion coding scheme and data relationship generation scheme are needed. Summary of the Invention

[0010] In view of this, the embodiments of this specification provide a spatiotemporal fusion coding method, a data relationship generation method and system, which solve the problem of how to uniformly identify and organize multi-source heterogeneous data. It can construct a unified spatiotemporal identity for multi-source heterogeneous data, and on this basis realize the automated association and relationship reasoning of cross-modal data, thereby breaking down data silos and providing basic support for large-scale spatiotemporal data analysis and intelligent applications.

[0011] The embodiments in this specification provide the following technical solutions: This specification provides a spatiotemporal fusion coding method, including: Acquire spatial, temporal, and object identification information of multi-source heterogeneous target data objects; Based on the geospatial grid coding system, the spatial information is mapped to a spatial code; wherein, the geospatial grid coding system forms a multi-level grid by recursively dividing the Earth's space, and each grid has a unique code; The time information is standardized, and based on a time-layered coding system, the standardized time information is mapped to a time code. Based on the object encoding engine, the attributes of the object identification information are inherited and transformed into standardized attribute information to generate a unique object identifier for the target data object; Based on a preset bit-level fusion rule, the spatial code, the temporal code, and the object identifier are associated to generate a spatiotemporal fusion code for the target data object, wherein the spatiotemporal fusion code is a structured binary code string that uniquely identifies the target data object; Based on the standardized attribute information, the spatiotemporal fusion encoding is bound to the attribute information of the target data object to obtain standardized data records, so as to construct a cross-system spatiotemporal relationship network.

[0012] This specification also provides a data relationship generation method, including: Obtain standardized data records for multiple target data objects, wherein each standardized data record is generated based on the aforementioned spatiotemporal fusion coding method, and the standardized data record includes at least: spatiotemporal fusion coding and attribute information bound to the spatiotemporal fusion coding; Based on the spatiotemporal fusion coding, the standardized data records are partitioned according to a preset level; Within each hierarchical partition, the spatiotemporal fusion encoding is decoded and / or logically operated on, and the spatiotemporal relationship between the attribute information of different data objects is derived based on the results of the decoding and / or logical operation. Based on the aforementioned spatiotemporal relationships, a data relationship network is constructed; Each spatiotemporal relationship data item in the data relationship network is associated with an evidence list and a confidence score. The evidence list records one or more standardized data records on which the spatiotemporal relationship is derived. The confidence score is obtained based on a preset scoring rule, which includes at least one of the following: the quality score of the data record in the evidence list, the strictness of the rule matching, the time decay factor, and the statistical score.

[0013] This specification also provides a spatiotemporal fusion coding system, including: The data acquisition module is used to acquire spatial information, temporal information, and object identification information of multi-source heterogeneous target data objects; The spatial coding module is used to map the spatial information into spatial codes based on a geospatial grid coding system; wherein the geospatial grid coding system forms a multi-level grid by recursively dividing the Earth's space, and each grid has a unique coding identifier; The time encoding module is used to standardize the time information and map the standardized time information into time codes based on a time-layered encoding system. The object identifier generation module is used to inherit the attributes of the object identifier information and convert them into standardized attribute information based on the object encoding engine, so as to generate a unique object identifier for the target data object; The fusion encoding generation module is used to establish an association between the spatial encoding, the temporal encoding and the object identifier based on a preset bit-level fusion rule, and generate a spatiotemporal fusion encoding of the target data object, wherein the spatiotemporal fusion encoding is a structured binary encoding string that uniquely identifies the target data object; The data record generation module is used to bind the spatiotemporal fusion encoding with the attribute information of the target data object based on the standardized attribute information to obtain standardized data records, so as to construct a cross-system spatiotemporal relationship network.

[0014] This specification also provides a data relationship generation system, including: The data acquisition module is used to acquire standardized data records of multiple target data objects, wherein each standardized data record is generated based on the above-mentioned spatiotemporal fusion coding method, and the standardized data record includes at least: spatiotemporal fusion coding and attribute information bound to the spatiotemporal fusion coding; The data partitioning module is used to partition the standardized data records according to a preset level based on the spatiotemporal fusion coding. The relationship deduction module is used to decode and / or perform logical operations on the spatiotemporal fusion encoding within each hierarchical partition, and deduce the spatiotemporal relationship between the attribute information of different data objects based on the results of the decoding and / or logical operation. A data relationship network construction module is used to construct a data relationship network based on the spatiotemporal relationship. Each spatiotemporal relationship data item in the data relationship network is associated with an evidence list and a confidence score. The evidence list records one or more standardized data records on which the spatiotemporal relationship is derived. The confidence score is obtained based on a preset scoring rule, which includes at least one of the following: the quality score of the data record in the evidence list, the strictness of the rule matching, the time decay factor, and the statistical score.

[0015] Compared with the prior art, the beneficial effects that at least one technical solution adopted in the embodiments of this specification can achieve include at least: This application assigns a unique, fixed-length, bit-operable spatiotemporal fusion code to any target data object by spatiotemporal fusion coding with fixed-length bit concatenation, and binds it with attribute information to form a standardized record. This breaks down the barriers between multi-source heterogeneity, realizes second-level indexing and relationship derivation of "one item, one code; one bit, one code", significantly reduces the complexity of cross-system association, and improves the efficiency of subsequent spatiotemporal relationship network construction and query accuracy. Attached Figure Description

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

[0017] Figure 1 This is a flowchart of a spatiotemporal fusion coding method in this application; Figure 2 This is a flowchart of a data relationship generation method in this application; Figure 3 This is the spatiotemporal fusion coding and relation generation logic diagram under the three-layer architecture in this application; Figure 4 This is a schematic diagram of the fields of the core database table in an embodiment of a data relationship generation method in this application; Figure 5 This is a data example diagram using spatiotemporal fusion coding technology in this application; Figure 6 It is a data processing flowchart; Figure 7 This is a flowchart of the raw data collection and processing process; Figure 8 It is the CAD design drawing before cleaning; Figure 9 These are the cleaned CAD design drawings; Figure 10 It is a vector file of the CAD design drawing; Figure 11 This is a schematic diagram of vectorized spatial data processing. Detailed Implementation

[0018] The embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0019] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. This application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] It should be noted that various aspects of embodiments within the scope of the appended claims are described below. It will be apparent that the aspects described herein can be embodied in a wide variety of forms, and any particular structure and / or function described herein is merely illustrative. Based on this application, those skilled in the art will understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number and aspects set forth herein can be used to implement the device and / or practice the method. Additionally, this device and / or method can be implemented using structures and / or functionalities other than one or more of the aspects set forth herein.

[0021] It should also be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. The drawings only show the components related to this application and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0022] Additionally, specific details are provided in the following description to facilitate a thorough understanding of the examples. However, those skilled in the art will understand that practice can be carried out without these specific details.

[0023] With the ever-increasing volume of data accumulating across various industries, existing technical solutions often employ universal identifiers such as DOIs and Handles to achieve globally unique data identification and resolution, or rely on spatial grid coding systems for data organization. For example, scientific data spaces use DOIs (such as 10.1109 / 5.771073) to associate metadata, and China Mobile's data network uses DSSNs (such as DSSN: 9a25c8187999465086c987757967c99f) to achieve cross-domain data routing. However, existing solutions have the following significant drawbacks: First, existing identifiers generally lack spatiotemporal attributes, can only provide logical positioning, and cannot be directly mapped to physical spatiotemporal coordinates, resulting in low efficiency of spatiotemporal association queries. Secondly, there is poor compatibility between different identification systems. For example, DOI and Handle, DSSN and internal business codes each have independent naming rules, parsing mechanisms and application scenarios, and lack unified mapping rules. This makes it impossible for data from different fields and departments to be seamlessly integrated based on unified identification, forming "data silos" and hindering information sharing across systems and industries.

[0024] In light of this, the inventors, through research and improvement exploration, discovered that although the current discrete grid coding system can effectively express spatial location (e.g., NO36J009300078083010010520430140), assign spatiotemporal labels to data, and support the construction of spatiotemporally driven data aggregation systems (such as the GDS grid aggregation system) and cross-center data routing, it lacks sufficient description of the time dimension. It typically requires additional fields and struggles to support efficient spatiotemporal integrated indexing and dynamic queries. In other words, the discrete grid coding system is more suitable for the discretization and identification of spatial locations, but it has limitations in describing non-spatial semantic relationships, dynamic evolutionary relationships, and cross-modal data associations between data objects.

[0025] In addition, due to the lack of a unified spatiotemporal unique ID and an attribute relationship construction mechanism based on the ID in the existing solutions, it is difficult for the industry to form an integrated panoramic view covering structured, unstructured and geographic information data, which greatly limits the in-depth mining and effective utilization of the value of the entire data chain in applications such as smart cities, asset management, disaster emergency response and intelligent operation and maintenance.

[0026] Based on this, the embodiments of this specification propose a spatiotemporal fusion coding method: The overall idea is as follows: spatial information of target data objects is mapped to a geospatial grid coding system to generate spatial codes; time information is standardized according to a preset time granularity (e.g., year, month, day, hour, etc.) and mapped to a hierarchical time coding system to generate time codes; then, spatial codes, time codes and object identification information are associated according to preset rules to generate unique spatiotemporal fusion codes; finally, the codes are bound to attribute information to form standardized data records, mapping data objects of different formats to a unified spatiotemporal coding system, laying the data foundation for building a unified spatiotemporal relationship network.

[0027] The technical solutions provided by the various embodiments of this application are described below with reference to the accompanying drawings.

[0028] like Figure 1 As shown in the embodiments of this specification, a spatiotemporal fusion coding method is provided, including: Step S100: Obtain the spatial information, temporal information, and object identification information of the multi-source heterogeneous target data object.

[0029] During implementation, an object coding technology system is introduced to assign an identity code to each physical entity and digital object, thereby forming a coding mechanism with a dual dimension of "time and space + object".

[0030] Specifically, the data access module collects raw data of objects from sources such as business databases, sensing devices, IoT terminals, remote sensing images, video surveillance, and government information systems. The collected objects are classified by type (such as natural objects, building facilities, equipment assets, behavioral events, personnel entities, virtual data objects, etc.). When each data object enters the system, a unique object registration information table is generated, which may contain the following fields: {Object Name, Object Type, Industry, Owner, Spatial Information, Time Information, Source System ID} Step S200: Based on the geospatial grid coding system, the spatial information is mapped to a spatial code; wherein, the geospatial grid coding system forms a multi-level grid by recursively dividing the Earth's space, and each grid has a unique code.

[0031] In implementation, the geospatial grid coding system adopts a recursive partitioning mechanism of "global latitude and longitude partitioning and quadtree hierarchy". Specifically, the partitioning rules are as follows: starting from the global level, it is recursively subdivided into four subgrids (quadtrees) according to the four equal parts. Each level deeper adds an additional quaternary number (0 / 1 / 2 / 3). The coding can go up to level 32 (sub-levels below the second); the coding format is commonly a quaternary one-dimensional code starting with "G" (G represents level 0, representing the entire Earth), and equivalent binary / decimal forms also exist for calculation and storage. The encoding length itself implies a hierarchy, with longer lengths containing finer details. The hierarchy is divided into four segments based on the granularity of the subdivision, corresponding to degrees, minutes, seconds, and units below the second in traditional latitude and longitude coordinates: Levels 1–9 are “degree-level grids”; Levels 10–15 are “hierarchical grids”; Levels 16–21 are “second-level grids”; and Levels 22–32 are “sub-second sub-grids”. Common documents often write it as: G ddddddddd-mmmmmm-ssssss.uuuuuuuuuuu (d / m / s / u are 0–3).

[0032] The geospatial grid coding system transforms real-world geospatial vector objects (including points, lines, and polygons) into computable grid codes through a unified grid division. Its mapping strategy can be determined based on the geometry type. Point encoding: Obtain a unique grid code (or a prefix code from level 0 to level). Start drilling down to Level 1 from the root (G): Within the current cell, determine the four quadrant sub-cell the point falls into based on its latitude / longitude relative to the cell center (or the cell's dividing line). Append the corresponding quaternary number (0 / 1 / 2 / 3) to obtain a quaternary string of length Level. Adding G to the prefix gives the point's encoding at that resolution.

[0033] Line encoding: Gets the set of grid codes (specified level) for all intersections / overlaps with line segments. A subdivided recursive strategy is adopted: recursively drill down from top to bottom. If the line does not intersect with the current cell, prune it; if it intersects and reaches the target level, collect the cell code; otherwise, continue drilling down four sub-cells.

[0034] Face encoding: Obtains the set of grid codes (specifying the level) for all faces that are covered or intersecting. A strategy combining recursive subdivision and relation testing is adopted: recursively from top to bottom. If the face is separate from the current cell, it is pruned; if it completely contains the current cell and can be directly "injected", it is collected in batches at the target level; if they intersect but have not reached the level, it is further subdivided to the level; if it reaches the level, the cell is collected.

[0035] This enables a unified mapping from geospatial space to cyberspace and then to data space, and constructs a cyber grid code (CGC) that can support virtual-real fusion computing and trusted identification.

[0036] Step S300: Standardize the time information and map the standardized time information into time codes based on the time hierarchical coding system.

[0037] In implementation, the time dimension may not be linear and needs to be processed according to time units (seconds / minutes / hours) or discrete levels. Specifically: First, convert all times into a unified time code / level (e.g., second-level timestamps plus optional hierarchical time slices). Then, standardize all time information based on UTC to eliminate the influence of time zones and daylight saving time, and establish a unified time benchmark. Finally, map the standardized time information to a mixed base hierarchical system of "year-month-day-hour-minute-second-microsecond" to generate a fixed-length binary time code.

[0038] This system organizes the timeline into a multi-level hierarchical structure from years to seconds, with each level representing a different temporal granularity, forming a "time precision level" corresponding to the spatial "resolution level". That is, each level is similar to a "resolution level" in space, and the finer the time level, the higher the precision, which is logically equivalent to the decomposition of time into a quadtree / decimal tree.

[0039] Shift right >> indicates moving from "seconds" to the parent level such as "minutes" or "hours".

[0040] Left shift << indicates moving down from "hours" to sub-levels such as "minutes" and "seconds".

[0041] Alternatively, a fixed-time-slice hierarchical structure (discrete-time grid) can be used, similar to a spatial grid, to divide the time axis into equal parts: 1 day --> 24-hour grid --> 60-minute grid --> 60-second grid Alternatively, you can customize the granularity: such as "one grid every 5 minutes" or "one grid every hour".

[0042] Step S400: Based on the object encoding engine, inherit the attributes of the object identification information and convert them into standardized attribute information to generate a unique object identifier for the target data object.

[0043] In implementation, the system calls the "object encoding engine" and adopts a hierarchical encoding structure, for example: OID = <Main Code> + <Object Type Code> + <Spatiotemporal Index Code> + <Serial Number> Entity code: Represents the entity or organization to which the data belongs; Object type code: such as "device=01", "video stream=02", "natural feature=03", etc.; Spatiotemporal index code: refers to the spatiotemporal grid code where the object is located; Serial number: An auto-incrementing number under the same subject + type + grid.

[0044] The system generates a unique object identifier (OID) for each object and writes it into a trusted registration table to achieve uniqueness, traceability, and verifiability.

[0045] Step S500: Based on a preset bit-level fusion rule, establish an association between the spatial code, the temporal code, and the object identifier to generate a spatiotemporal fusion code for the target data object, wherein the spatiotemporal fusion code is a structured binary code string that uniquely identifies the target data object.

[0046] Specifically, information in the three dimensions of space, spatiotemporal, and objects is encapsulated into a "standardized digital product" that a computer can directly understand and process efficiently, so that all subsequent complex spatiotemporal relationship reasoning can be transformed into extremely fast bit operations.

[0047] Step S600: Based on the standardized attribute information, bind the spatiotemporal fusion code with the attribute information of the target data object to obtain standardized data records, so as to construct a cross-system spatiotemporal relationship network.

[0048] In implementation, by binding the generated spatiotemporal fusion code with the attribute information of the target data object, a unified standardized data record is formed, providing a foundation for the subsequent construction of a spatiotemporal relationship network. This record, with the spatiotemporal fusion code as the core index, achieves integrated anchoring of object identity, spatiotemporal location, and business attributes. It provides a unique and reliable data foundation for automatically deriving the spatiotemporal relationships and semantic associations between objects based on encoding operations, thereby supporting the construction of a computable and traceable spatiotemporal relationship network.

[0049] It's important to clarify that spatiotemporal fusion coding is not simply "spatial coding + temporal coding." It refers to a big data technology system that follows a "spatiotemporal-based, object-integrated" approach, closely combining spatiotemporal grid coding with object information coding to form an integrated coding and application system encompassing "spatiotemporal location + object information." This system covers functions such as coding identification, indexing, representation, and computation. In simple terms, it treats the grid as containers arranged by location, placing various types of data inside according to their location, and then displaying, analyzing, and serving them according to their container numbers.

[0050] This application establishes a correlation and fusion relationship based on the spatial and state changes of objects through spatiotemporal grid coding, thereby realizing the transformation from a multi-dimensional heterogeneous data structure (object description) to a uni-dimensional homogeneous data structure (spatiotemporal description).

[0051] Spatiotemporal fusion coding can be extended to the virtual network world, establishing a CyberGrid Code (CGC) technology system composed of three dimensions: geographic location, IP address, and trustworthiness. This system identifies the network space units where data resources reside and assigns a fusion code of "spatiotemporal + object" to the content of data objects, innovatively solving the fundamental problems of organizing big data in cyberspace and integrating virtual and physical spaces. The basic framework of this technology is a spatiotemporal big data grid foundation that integrates "geospatial space—cyberspace—data space." Its core idea is to perform hierarchical subdivision and ordered coding of multi-dimensional space (geographic location, network address, data object) under a unified grid coding system, realizing the organization, mapping, and computation of data resources across spatial scenarios.

[0052] In some embodiments, before the association is performed based on a preset bit-level fusion rule, the method further includes: Establish a mapping relationship between the object identifier and the spatial code; If the target data object is a moving object, then based on the location code of the target data object, the corresponding spatiotemporal code is dynamically updated in units of a preset timestamp, and the mapping relationship between the object identifier and the spatiotemporal code is maintained to form a spatiotemporal trajectory table.

[0053] In conjunction with the above embodiments, the spatiotemporal binding module can be used to associate an object identifier (OID) with its corresponding spatiotemporal grid code (GID): Mapping: OID GID If the object is mobile (such as a vehicle, person, drone, etc.), the system generates a new spatiotemporal index each time its location is updated, forming a temporal trajectory table of the object: {OID, GID_t1, GID_t2, … GID_tn} This allows for the construction of a dynamic spatiotemporal-object relationship chain.

[0054] In some embodiments, all OID registration information is written to a Trusted Code Registry. The system provides an "encoding parsing service" that supports: Forward resolution (OID --> object properties + spatiotemporal location) Reverse parsing (GID + condition --> list of matching objects) This mechanism ensures that objects in different systems can be mutually recognized, queried, and computed within the same spatial reference frame.

[0055] In some embodiments, the preset bit-level fusion rules include: bit splicing rules or interleaving rules.

[0056] In implementation, if a concatenation rule is adopted, the three encoded parts are directly concatenated in a predetermined order, like strings, for example: Spatial code 101001 (representing a specific land parcel) The time code 202509111435 (represents September 11, 2025, 14:35) Object Identifier (OID): BJ_Device_12345 The merged encoding is 101001202509111435BJ_Device_12345. Alternatively, the binary strings of the three can be directly concatenated.

[0057] If an interleaving rule is adopted, the binary bits of the three are interleaved and mixed together according to a specific algorithm, such as using Morton code (also known as Z-order curve) or a three-dimensional interleaving (x, y, t) method, which interleaves the X and Y bits of the spatial coordinates to generate a fused code.

[0058] In some embodiments, the target data object includes: structured and unstructured data oriented towards the physical world, wherein the unstructured data includes: geographic information data; If the target data object is structured data, then the spatial information, the time information, and the object identification information are obtained by parsing the database fields corresponding to the target data object; If the target data object is unstructured data, the spatial information and the temporal information are obtained from the content of the target data object through information extraction, and the object identification information is obtained by parsing the file metadata, business number or system allocation of the target data object. If the target data object is geographic information data, then the internal coordinates and dataset metadata of the target data object are parsed to obtain the spatial information and the temporal information, and the object identification information is obtained by parsing the feature identifier of the target data object; wherein, the geographic information data includes at least one of the following: Building Information Model (BIM), Computer-Aided Design (CAD) documents, remote sensing imagery or vector layer.

[0059] This application integrates spatial location coding and temporal dimension coding into a unified framework, supporting the processing of linear and nonlinear time units (seconds, minutes, hours, discrete levels); and realizes cross-level and cross-dimensional data mapping based on the partitioned grid to form a unified spatiotemporal ID.

[0060] This application binds multi-source heterogeneous data, such as structured data and unstructured data (including geographic information data) facing the physical world, with a unique spatiotemporal fusion code to achieve unified anchoring of object attributes and spatiotemporal location / time.

[0061] Based on the same inventive concept, such as Figure 2 As shown, this application also provides a data relationship generation method, including: Obtain standardized data records for multiple target data objects, wherein each standardized data record is generated based on the aforementioned spatiotemporal fusion coding method, and the standardized data record includes at least: spatiotemporal fusion coding and attribute information bound to the spatiotemporal fusion coding; Based on the spatiotemporal fusion coding, the standardized data records are partitioned according to a preset level; Within each hierarchical partition, the spatiotemporal fusion encoding is decoded and / or logically operated on, and the spatiotemporal relationship between the attribute information of different data objects is derived based on the results of the decoding and / or logical operation. Based on the aforementioned spatiotemporal relationships, a data relationship network is constructed; Each spatiotemporal relationship data item in the data relationship network is associated with an evidence list and a confidence score. The evidence list records one or more standardized data records on which the spatiotemporal relationship is derived. The confidence score is obtained based on a preset scoring rule, which includes at least one of the following: the quality score of the data record in the evidence list, the strictness of the rule matching, the time decay factor, and the statistical score.

[0062] In practice, each raw data item is transformed into a standardized data record (i.e., a standard unit Record) at the data organization layer: Record = {object_id, original_id, grid_code, time_code, fused_code,attributes: {k:v,...}, source, quality_score, ts_utc} Here, object_id represents the system's unique object ID. If the original data already has a unique identifier (device serial number, sensor ID + timestamp), then object_id is generated using hashing; otherwise, the system generates a UUID. original_id represents the original data source ID, which can be directly taken as the unique identifier of the original data (such as device serial number, database primary key). grid_code represents spatial grid encoding, which maps geographic coordinates to grid cells of a preset resolution. Line / polygon objects are first subdivided into sub-objects and then mapped to the overlay grid. time_code represents time encoding, which converts the original time into a UTC timestamp or hierarchical time encoding (such as YYYYMMDDHHMMSS). fused_code represents spatiotemporal fusion coding, which merges object_id, grid_code and time_code to generate a hash or cyber grid code (CGC). attributes represent a set of business attributes, which extract the core features of the original data and convert them into key-value pair format; "Source" indicates the data origin, identifying the original acquisition system, equipment, sensor, or data provider. The quality_score represents the data quality score (source credibility / location accuracy), and is generated as a value between 0 and 1 based on indicators such as completeness, accuracy, and anomaly detection. ts_utc represents the entry time, which is generally a standardized timestamp (UTC).

[0063] First, obtain standardized data records containing spatiotemporal fusion codes (fused_code) and attribute information to ensure that subsequent processing can be directly performed based on the coded bit segments.

[0064] To improve the processing efficiency of massive amounts of data, a partition key is generated based on the fixed bit segment representing a specific level in the fused_code. Data records are then divided into buckets of different levels without overlap, achieving consistency in the encoding structure within the same bucket. This partitioning process is completed in a distributed computing framework. The <level, record> key-value pairs are directly output through the Map phase, and parallel processing is performed by bucket in the Reduce phase, thereby effectively avoiding full table scans and significantly improving processing efficiency.

[0065] Within each level of partition, the system performs decoding operations and logical operations on the spatiotemporal fusion encoding, automatically derives the spatiotemporal relationships between the attributes of different data objects based on the operation results, and finally integrates all the derived spatiotemporal relationships to construct a complete data relationship network.

[0066] Each relational data item in the data relational network of this application does not exist in isolation, but is associated with an evidence list (used to trace and deduce the original data records on which the relation is based) and a confidence score (calculated by comprehensively considering factors such as data quality, rule strictness, and time decay), thereby forming a traceable, quantifiable, and trustworthy spatiotemporal relational graph.

[0067] The goal of this application is to map "attributes (fields or meta-information of any source and modality)" onto spatiotemporal units through spatiotemporal fusion encoding, and then automatically derive the spatiotemporal relationships between attributes (co-occurrence, proximity, evolution / sequence, coverage / aggregation, intersection / union, correlation / causal candidate, etc.) according to rules, arithmetic and bit operations, and statistical / learning methods, and assign evidence and confidence to each relationship to form a queryable and updatable spatiotemporal relationship network.

[0068] In some embodiments, the confidence level and scoring formula are as follows: score(rel) = normalize( w_rule*rule_score + w_stat*stat_score + w_src*src_score + w_time*temporal_decay ) Here, rule_score represents the strictness of rule matching (e.g., same position = 1, nearest neighbor = 0.8). stat_score indicates that if the relationship is statistical, the absolute value of the correlation coefficient or normalized mutual information is used; src_score represents the combination of source credibility (e.g., min(qA,qB) or weighted average); temporal_decay = exp(- (now - evidence_time) / τ): Decays the influence of old evidence over time.

[0069] All weights are configured by the system or learned from labeled data.

[0070] In some embodiments, for non-spatial native data (documents, images), location information should be extracted first (NER, OCR, BIM parsing, image landmark recognition), and a location confidence score (quality_score) should be attached.

[0071] In some embodiments, multiple fault-tolerance and adaptive mechanisms are introduced to effectively handle the heterogeneity and uncertainty in the data fusion process: Non-spatial native data: Location extraction is performed and confidence scores are assigned; if the confidence score is low, strong relationships are not generated, only "candidate relationships" are generated. Different time granularities: unify to the smallest common granularity or retain multiple levels (second / minute / hour level) at the same time, and use hierarchical mapping to handle cross-granularity matching (using >> / mask); Conflict fusion: If there is evidence of conflict within the same relationship, record all evidence and use scoring rules to determine the final state (accepted / candidate / rejected). Thresholds and human review: The system automatically generates candidate relationships and provides a channel for manual review (especially causal / high-value relationships).

[0072] It should be noted that, as Figure 3 As shown, the spatiotemporal big data grid base consists of three parts: the physical layer, the logical layer, and the conceptual layer.

[0073] Physical Layer: The implementation of the foundation uses grid coding as the core indexing mechanism, mapping spatial objects such as points, lines, and surfaces to unified grid cells. Each object is bound to attribute information through a grid coding ID, achieving object-level unique identification and traceability. For example, in the standardized data record formed by this binding, Cell_ID represents the spatial location and is the smallest unit after gridding, allowing the physical location of each object to be accurately located through the Cell ID. The Attributes field is a collection of attribute information of objects within the Cell, used to store the static or dynamic attributes of the objects, such as category, area, purpose, ownership, etc. The Objects field is a list of specific objects within the Cell, each with a unique ID, thereby mapping raw data objects from different sources and in different formats to the unified format of "standardized data records" under a unified coding system. This layer provides the basic capabilities for spatiotemporal fusion coding and object data association.

[0074] The logical layer, building upon the physical layer encoding, achieves hierarchical representation and relative computation of objects by partitioning entity objects and their adjacency relationships. Typical operations include hierarchical representation of points, lines, and surfaces (VPoint, VLine, VArea), sub-object decomposition (SubPoint, SubLine, SubArea), and adjacency and topological relationship calculations within the partitioned grid (RPoint, RLine, RArea). It defines the process of object representation from vector representation to partitioning to rasterization. This layer solves the problem of unifying cross-scale representation and multi-granularity computation, providing a mechanism for refined modeling and relational deduction of large-scale complex objects.

[0075] Conceptual Layer: Supported by the logical layer relationships, a fusion expression model of "spatiotemporal + object" is constructed. Specifically, the encoding function O(Level) maps geographic coordinates (Lat, Lon), network addresses (IP), and data object codes (UCode) to a unified grid cell (UPixel). O(Level) represents how objects are encoded at a specific resolution / level; ∪ {BCode, UCode} represents converting an object into one or more grid codes (BCode can be understood as the basic grid code, and UCode is the object's unique code). The content focuses on how points, lines, and surfaces in the real world (geographic space) are transformed into computable codes through a unified grid division. This layer achieves a unified mapping from geographic space to cyberspace and then to data space, constructing a cyber-grid code (CGC) that supports virtual-real fusion computing and trusted identification.

[0076] The complete spatial encoding formula for an object is as follows:

[0077] Wherein, GridID is the spatial code; TimeID is the time code; ObjectID is the object code; m and n represent the spatial level number; and i represents the index.

[0078] In some embodiments, the logical operation includes at least one of bitwise AND, bitwise OR, bitwise XOR, bitwise NOT, and shift operation; The bitwise AND operation is used to extract the common parent grid or parent time interval; the bitwise OR operation is used to construct the minimum union of covering or approximate sets; the bitwise XOR operation is used to assist in determining proximity relationships; the bitwise NOT operation is used to exclude specific spatiotemporal units or perform reverse selection; and the bitwise shift operation is used to perform conversion between different levels to obtain parent or child level codes.

[0079] In practice, bitwise operations combined with hierarchical shifting can efficiently determine parent-child / sibling relationships (>> / <<), common parent regions (AND mask), and unions (OR). However, a more robust approach for neighborhood determination is to decode the grid coordinates (x, y[, z]) and time index t, and then use the distance or index difference to make the determination.

[0080] In spatiotemporal fusion coding, logical operations are performed directly on the binary bits of the code, enabling bit-level operations to map common spatial / temporal relationships: Bitwise AND preserves the high-order common prefix, equivalent to extracting the nearest common parent grid or parent time interval between two codes (e.g., A=101011, B=101001, A&B=101000 represents a common parent grid); Bitwise OR can merge bit information for quickly constructing a minimum union of coverage or approximation; Bitwise XOR reveals differences in bits, and the number and position of the differences can be used to determine whether they are adjacent grids or difference patterns (e.g., only the least significant bit differs, usually indicating adjacent grids); Bitwise NOT generates a complement mask for exclusion or reverse selection; Left / Right shift (<>) is a hierarchical operation—right shift removes lower-level details to obtain the parent layer code (coarsening the granularity), left shift adds lower-order bits for drilling down to sub-layers or restoring bit positions in interleaved coding; if interleaved (Mor) is used... In a ton / Z-order or three-dimensional interleaving (x,y,t) scheme, temporal bits can participate in these bit operations just like spatial bits, thereby enabling the simultaneous determination of composite relationships of "being in the same parent interval at the same time and in adjacent spatial grids". It should be noted that the premise for these bit operations to be valid is a unified bit layout and hierarchy / bitmask strategy (e.g., fixed hierarchy length, padding / alignment rules), and they are often combined with masks and shifts to extract bits of a certain level or dimension, thereby completing coverage, intersection, neighborhood, difference and hierarchy queries in massive spatiotemporal indexes with extremely low computational cost.

[0081] It should be noted that the addition, subtraction, difference, multiplication and division operations in the encoding need to take into account both space and time dimensions.

[0082] In some embodiments, the derivation of the spatiotemporal relationship between the attribute information of different data objects includes: The spatiotemporal fusion codes are matched based on deterministic rules to generate basic spatiotemporal relationships; wherein, the deterministic rules include comparing the matrix unit values ​​between spatiotemporal fusion codes through a matrix-based encoding data structure to determine the basic spatiotemporal relationships between objects; Furthermore, statistical or sequence analysis is performed on the attribute information based on probabilistic rules to generate statistical spatiotemporal relationships; wherein, adjacent, disjoint, and merged relationships constitute a basic spatiotemporal relationship set; the basic relationships serve as the basis for inferring complex spatiotemporal relationships to expand and generate complex spatiotemporal relationship types, wherein the complex spatiotemporal relationship types include at least one of the following: covering relationship, inclusion relationship, co-position relationship, intersection / union relationship; the statistical spatiotemporal relationships include at least one of the following: evolutionary / sequence relationship and correlation relationship.

[0083] In practice, attributes occurring within the same grid or the same grid level are more likely to generate "co-occurrence relationships".

[0084] Specifically, the following are some commonly used relationship types and decision rules (pseudocode / decision logic): 1. Co-location / Co-occurrence Definition: Two attributes are in the same time-space grid cell and their time windows overlap (or in the same time slice).

[0085] Decision: if recordA.fused_code[levelL] == recordB.fused_code[levelL]andtime_overlap(recordA, recordB, Δt) → generate CO_OCCUR.

[0086] time_overlap: can be the same time slice, or |ts_a - ts_b| <= Δt_threshold (e.g., Δt=0 means the same second, or Δt=300s means a 5-minute window).

[0087] Evidence: original_id, source, time difference, and quality_score of the two records.

[0088] Confidence level (example formula): conf = w_loc*min(qA,qB) * w_time*exp(-|Δt| / τ) (w_loc, w_time, τ are parameter tuning constants).

[0089] 2. Spatial / Temporal Neighbor: XOR operation on each bit. Definition: The attribute is located in an adjacent grid or adjacent time slice (spatially typically a Manhattan / Chebyshev distance of 1).

[0090] Determination: The safest approach: After decoding (xA,yA,tA) = decode(fused_codeA) and (xB,yB,tB), if spatial_dist((xA,yA),(xB,yB)) ≤ d_s and |tA - tB| ≤ d_t, then NEIGHBOR is possible.

[0091] Bitwise operation aid (fast): If Morton encoding is used, you can first XOR = codeA ^ codeB. If hamming_weight(XOR) is very small or XOR is in the low bit range, they may be adjacent - but decoding verification is required to avoid misjudgment.

[0092] Confidence combines spatial accuracy, time window, and source quality.

[0093] 3. Coverage / Parent-Child Relationship (Aggregation / Containment) (Bitwise AND) Definition: An attribute falls within the parent grid or time parent interval of another attribute (hierarchical aggregation).

[0094] Determination: if (parent_code = fused_codeA>>k) == fused_codeB → CONTAINS / IS_CHILD_OF.

[0095] Here, >>k is an operation of hierarchical shift or clearing low bits by mask (depending on the coding implementation). A mask can also be used: if (codeA&mask_levelL) == codeB.

[0096] Usage: For roll-up aggregation (statistics, summarization).

[0097] 4. Temporal Evolution / Sequence Relationship Definition: Ordered attribute changes are formed over time for the same object or the same spatial unit (temporal sequence association).

[0098] Determination: Sort the time series t0<t1<... for the same object_id or the same grid_cell. If a recognizable pattern (rising, falling, mutation) occurs in value(t) over time, generate an EVOLVES_TO or TRANSITION relationship, recording from_attr, to_attr, Δt, and magnitude.

[0099] Implementation method: Sliding window + statistical tests (CUSUM, change point detection) or use models (ARIMA, LSTM) to detect significant changes and output event edges.

[0100] 5. Intersection / Union / Difference (Set relations) Definition: Synthesize the coverage of multiple attributes or find differences through bitwise operations. [[ID=?]]

[0101] Determination: OR combines regions, AND finds the common parent region, XOR finds different bits. Usually used for "the smallest grid set covered by which attributes simultaneously".

[0102] 6. Correlation / Association Relationship Definition: Two attributes exhibit statistical correlation (positive / negative correlation), frequent co-occurrence, or potential causal candidates on the same spatiotemporal network set.

[0103] Judgment method: Calculate the correlation index (Pearson, Spearman, mutual information, lift / support / confidence, etc.) of attribute pairs on multiple samples for each grid cell or time window. If it exceeds the threshold, generate ASSOCIATED. You can further use Granger causality test or causal graph model to label POTENTIAL_CAUSAL (Note: label as "candidate causality" and need to be manually / experimentally verified).

[0104] In some embodiments, the data relationship generation method further includes: relationship lifecycle management; When a new standardized data record is added to the database, perform the following operations: If the standardized data record is associated with an existing spatiotemporal relationship, the standardized data record is added as evidence to the evidence list of the spatiotemporal relationship, and the confidence score of the spatiotemporal relationship is recalculated based on the updated evidence list. Otherwise, based on the standardized data records and other existing records, the relationship is deduced through the deterministic or probabilistic rules. If a spatiotemporal relationship that does not yet exist in the data relationship network is deduced, a corresponding new spatiotemporal relationship data item is generated, and an initial evidence list and confidence score are created. Based on preset failure rules, the spatiotemporal relationships in the data relationship network are periodically checked, and spatiotemporal relationships with confidence scores below preset thresholds or without new evidence within preset time periods are marked as failures or archived. Preserve the complete chain of evidence and data source information for all spatiotemporal relationships to support historical retrospection and human intervention correction of the relationship derivation process.

[0105] During implementation, the process includes: creation: rules or models are discovered and written to the relation database (status: candidate / confirmed); update: when new evidence comes in, evidence is added and the score is recalculated; decay / deletion: relations that have not received new evidence for a long time and whose score is below the threshold are automatically expired; audit / rollback: the evidence chain and source are preserved, supporting backtracking and manual intervention.

[0106] In some implementations, MapReduce / Spark is used: the Map phase is partitioned by fused_code@level, and the Reduce phase performs neighborhood / same-position determination within the grid bucket.

[0107] In some embodiments, the neighborhood table and mask are pre-computed to reduce decoding costs.

[0108] In some implementations, hotspot grids (high-frequency update areas) are processed in real-time stream (Kafka + Flink), while static batch processing is used for offline statistical mining.

[0109] In some embodiments, spatial sharding and replication strategies are used for large-scale spaces.

[0110] In some embodiments, multi-dimensional result presentation and verification mechanisms are provided. Specifically, spatiotemporal heatmaps, trajectory animations, and attribute co-occurrence networks are used to visualize and intuitively display the analysis results; a classification confusion matrix is ​​used to verify deterministic relationships, and cross-validation and confidence interval estimation are used to assess the reliability of statistical relationships; a feedback closed-loop system is established to feed back manual annotations and feedback from the business system to the model layer, continuously optimizing parameter configurations and improving relationship confidence, thereby achieving self-evolution and continuous improvement of system performance.

[0111] This application uses arithmetic operations (addition, subtraction, difference, multiplication, division) and logical operations (AND, OR, XOR, shift, etc.) to quickly generate relationships between attributes at the encoding level, supporting automatic derivation of co-position relationships, proximity relationships, evolutionary relationships, covering relationships, and intersection / union relationships.

[0112] This application maps data objects from different sources and in different formats to a unified spatiotemporal ID, breaking down data silos and generating a computable and verifiable spatiotemporal relationship network on this basis, providing support for subsequent AI analysis and applications.

[0113] Below is another example of a data relationship generation method, which combines the previous examples to form an illustrative example.

[0114] Spatiotemporal coding inherently possesses temporal and spatial information. Based on bitwise operations within the coding itself, spatial relationships between attributes can be derived. In implementation, the coding engine system performs structured processing on the CAD data: each spatial element (such as polygons, lines, and points) in the vector file is extracted, converted into an internally structured record, and attribute fields are extracted. For each spatial element, a unique spatiotemporal code is generated based on parameters such as the element's geometry, floor level, and coding level. After processing, the structured spatial codes and attribute data are batch-written into the database.

[0115] like Figure 4 As shown, three core tables are created in the database: The spatial coding table (stores spatial spatiotemporal codes and their levels, geometric information, etc.), as shown on the left: code: stores the unique spatial spatiotemporal code generated for each spatial element and its level, geometric range, coordinates and other information. This is the spatial index table of all spatial objects and is the core anchor point of the system, used to indicate "where". The attribute table (storing the attribute fields of each spatial element) The "attributes" module on the right side of the diagram stores the business attribute fields of each spatial element (such as name, purpose, floor, area, category, etc.), representing "what" and decoupled from the coded content for easy independent maintenance; Registration parsing table (correspondence between associated spatial codes and attributes); The line in the middle of the diagram represents the "code_attribute_mapping" module; The correspondence between storage space encoding and attributes (feature_id) spatial_time_code); It is a bridge table (Mapping Table) between the two, used for registration, parsing, and related retrieval.

[0116] The following example uses the XX Building: After importing the building's CAD file, extract each spatial element: Room boundaries (polygons), walls and corridors (line elements), and cameras (point elements); Calculate the geometric center (Lat, Lon) and level (such as floor number, spatial resolution) of each element, generate a unique code (spatial_time_code) and write it into the code table; Then read the attribute information from the external file and write it into the attribute table, for example: Element: Meeting Room A; Attribute fields: Name = Meeting Room A; Area = 120㎡; Purpose = Meeting; Floor = 10; Source: CAD layer properties.

[0117] Then the system automatically binds the spatial code with the attribute record to generate, such as Figure 5 The registration resolution table shown (before registration resolution).

[0118] Then perform relation calculations based on the three tables: For example Meeting Room A Meeting room B has two adjacent codes with different low-order bits, belonging to adjacent grids; Meeting Room A Smoke detector 1 has overlapping coordinates and the same grid encoding; Correspondingly, the process of this embodiment is as follows: Figure 6 As shown, taking the xx building case as an example, the original data collection and organization includes, for example: Financial data: Construction in progress, 2 data entries; Archived data: CAD data, data size 0.0339GB; Processing method: Clean the original data and perform vector conversion; Manual + automated tools; The process of raw data collection and processing is as follows: Figure 7 As shown, the CAD electronic design data is first cleaned using a combination of automated general data processing and manual processing of special data to obtain the required architectural CAD data. Figure 8 and Figure 9 The comparison of data states before and after cleaning is shown. Subsequently, the cleaned CAD design drawings are converted into vector data with latitude and longitude coordinates, forming... Figure 10 The file shown is a computable vector data file.

[0119] Next, the transformed data is standardized through a combination of automated general data processing and manual processing of specific data, generating data such as... Figure 11 The output shows standard-compliant vector data including multiple attribute fields such as building, region, name, floor, area, and category, and is output in geojson file format.

[0120] Finally, data quality inspection is conducted to check the completeness of the processed data's related attribute information and coordinate information, and to compare it with the original CAD data to verify accuracy. After passing the quality inspection, the data is packaged in geojson format and WGS84 coordinate system as the standard form and submitted to the subsequent processing stage.

[0121] Based on the same inventive concept, this application also provides a spatiotemporal fusion coding system, comprising: The data acquisition module is used to acquire spatial information, temporal information, and object identification information of multi-source heterogeneous target data objects; The spatial coding module is used to map the spatial information into spatial codes based on a geospatial grid coding system; wherein the geospatial grid coding system forms a multi-level grid by recursively dividing the Earth's space, and each grid has a unique coding identifier; The time encoding module is used to standardize the time information and map the standardized time information into time codes based on a time-layered encoding system. The object identifier generation module is used to inherit the attributes of the object identifier information and convert them into standardized attribute information based on the object encoding engine, so as to generate a unique object identifier for the target data object; The fusion encoding generation module is used to establish an association between the spatial encoding, the temporal encoding and the object identifier based on a preset bit-level fusion rule, and generate a spatiotemporal fusion encoding of the target data object, wherein the spatiotemporal fusion encoding is a structured binary encoding string that uniquely identifies the target data object; The data record generation module is used to bind the spatiotemporal fusion encoding with the attribute information of the target data object based on the standardized attribute information to obtain standardized data records, so as to construct a cross-system spatiotemporal relationship network.

[0122] Based on the same inventive concept, this application also provides a data relationship generation system, comprising: The data acquisition module is used to acquire standardized data records of multiple target data objects, wherein each standardized data record is generated based on the spatiotemporal fusion coding method described above, and the standardized data record includes at least: spatiotemporal fusion coding and attribute information bound to the spatiotemporal fusion coding; The data partitioning module is used to partition the standardized data records according to a preset level based on the spatiotemporal fusion coding. The relationship deduction module is used to decode and / or perform logical operations on the spatiotemporal fusion encoding within each hierarchical partition, and deduce the spatiotemporal relationship between the attribute information of different data objects based on the results of the decoding and / or logical operation. A data relationship network construction module is used to construct a data relationship network based on the spatiotemporal relationship. Each spatiotemporal relationship data item in the data relationship network is associated with an evidence list and a confidence score. The evidence list records one or more standardized data records on which the spatiotemporal relationship is derived. The confidence score is obtained based on a preset scoring rule, which includes at least one of the following: the quality score of the data record in the evidence list, the strictness of the rule matching, the time decay factor, and the statistical score.

[0123] This application, through a spatiotemporal big data grid foundation, enables spatiotemporal fusion coding technology to simultaneously support: A globally unified, seamless identification system that features "one code per person," "one code per item," and "one code per event." Computationally efficient: Facilitates rapid encoding and recognition by computers, adapting to the development of big data technology and future unmanned technology; Precise positioning: It has precise location indication capabilities and unified spatial addressing of data; Seamless inheritance: Adaptable to applications, it extends existing systems or coding standards, has low implementation costs, and offers huge potential for expansion. Two-dimensional and three-dimensional integration: It simultaneously supports the generation and identification of spatiotemporal three-dimensional mesh maps and block data on the spatiotemporal cloud platform, facilitating collaboration and data integration between different business systems and businesses; Intelligent empowerment: Using fusion coding as the anchor point and combining knowledge graphs to carry out spatiotemporal intelligent computing, it injects automatic reasoning and decision-making capabilities into various business systems.

[0124] This invention proposes a spatiotemporal fusion coding method and an attribute relationship generation mechanism based on this coding, achieving unified identification and relationship calculation for structured and unstructured data (including geographic information data) across the physical world. Compared with existing technologies, this invention has the following advantages: First, it overcomes the limitations of traditional identifier systems, which can only uniquely identify objects and cannot generate relationships between them. It can automatically deduce various relationships between attributes, such as co-location, proximity, evolution, and coverage, under a unified spatiotemporal ID system. Second, compared with single spatial discretization coding methods, this invention can simultaneously handle both spatial and temporal dimensions, making it particularly suitable for processing nonlinear time series or discrete levels, thus meeting the data management and computation needs in complex dynamic scenarios. Third, this invention utilizes arithmetic and logical operations to achieve rapid calculation of spatiotemporal relationships, significantly improving the efficiency of querying, analyzing, and fusing multi-source heterogeneous data. Fourth, this invention provides a unified technical foundation for the confirmation, circulation, and trusted computation of data elements, supporting large-scale applications in emerging fields such as smart cities, industrial internet, spatiotemporal search, and low-altitude economy. In summary, this invention not only improves the accuracy and efficiency of spatiotemporal data organization and computation, but also lays a solid technical foundation for the value mining and asset utilization of multi-source heterogeneous data.

[0125] In this specification, the same or similar parts between the various embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the descriptions of the embodiments described later are relatively simple, and relevant parts can be referred to the descriptions of the foregoing embodiments.

[0126] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A spatiotemporal fusion coding method, characterized in that, include: Acquire spatial, temporal, and object identification information of multi-source heterogeneous target data objects; Based on the geospatial grid coding system, the spatial information is mapped to a spatial code; wherein, the geospatial grid coding system forms a multi-level grid by recursively dividing the Earth's space, and each grid has a unique code; The time information is standardized, and based on a time-layered coding system, the standardized time information is mapped to a time code. Based on the object encoding engine, the attributes of the object identification information are inherited and transformed into standardized attribute information to generate a unique object identifier for the target data object; Based on a preset bit-level fusion rule, the spatial code, the temporal code, and the object identifier are associated to generate a spatiotemporal fusion code for the target data object, wherein the spatiotemporal fusion code is a structured binary code string that uniquely identifies the target data object; Based on the standardized attribute information, the spatiotemporal fusion encoding is bound to the attribute information of the target data object to obtain standardized data records, so as to construct a cross-system spatiotemporal relationship network.

2. The spatiotemporal fusion coding method according to claim 1, characterized in that, Before the association based on the preset bit-level fusion rules, the following is also included: Establish a mapping relationship between the object identifier and the spatial code; If the target data object is a moving object, then according to the location code of the target data object, the corresponding spatiotemporal code is dynamically updated in units of a preset timestamp, and the mapping relationship between the object identifier and the time code is maintained to form a spatiotemporal trajectory table.

3. The spatiotemporal fusion coding method according to claim 1, characterized in that, The preset bit-level fusion rules include: bit splicing rules or interleaving rules.

4. The spatiotemporal fusion coding method according to claim 1, characterized in that, The target data objects include: structured and unstructured data oriented towards the physical world, wherein the unstructured data includes: geographic information data; If the target data object is structured data, then the spatial information, the time information, and the object identification information are obtained by parsing the database fields corresponding to the target data object; If the target data object is unstructured data, the spatial information and the temporal information are obtained from the content of the target data object through natural language processing, and the object identification information is obtained by parsing the file metadata, business number or system allocation of the target data object. If the target data object is geographic information data, then the internal coordinates and dataset metadata of the target data object are parsed to obtain the spatial information and the temporal information, and the object identification information is obtained by parsing the feature identifier of the target data object; wherein, the geographic information data includes at least one of the following: building information model, computer-aided design document, remote sensing image or vector layer.

5. A method for generating data relationships, characterized in that, include: Obtain standardized data records for multiple target data objects, wherein each standardized data record is generated based on the spatiotemporal fusion coding method described in any one of claims 1-4, and the standardized data record includes at least: spatiotemporal fusion coding and attribute information bound to the spatiotemporal fusion coding; Based on the spatiotemporal fusion coding, the standardized data records are partitioned according to a preset level; Within each hierarchical partition, the spatiotemporal fusion encoding is decoded and / or logically operated on, and the spatiotemporal relationship between the attribute information of different data objects is derived based on the results of the decoding and / or logical operation. Based on the aforementioned spatiotemporal relationships, a data relationship network is constructed; Each spatiotemporal relationship data item in the data relationship network is associated with an evidence list and a confidence score. The evidence list records one or more standardized data records on which the spatiotemporal relationship is derived. The confidence score is obtained based on a preset scoring rule, which includes at least one of the following: the quality score of the data record in the evidence list, the strictness of the rule matching, the time decay factor, and the statistical score.

6. The data relationship generation method according to claim 5, characterized in that, The logical operations include at least one of bitwise AND, bitwise OR, bitwise XOR, bitwise NOT, and shift operations; The bitwise AND operation is used to extract the common parent grid or parent time interval; the bitwise OR operation is used to construct the minimum union of covering or approximate sets; the bitwise XOR operation is used to assist in determining proximity relationships; the bitwise NOT operation is used to exclude specific spatiotemporal units or perform reverse selection; and the bitwise shift operation is used to perform conversion between different levels to obtain parent or child level codes.

7. The data relationship generation method according to claim 5, characterized in that, The derivation of the spatiotemporal relationships between the attribute information of different data objects includes: The spatiotemporal fusion codes are matched based on deterministic rules to generate basic spatiotemporal relationships; wherein, the deterministic rules include comparing the matrix unit values ​​between spatiotemporal fusion codes through a matrix-based encoding data structure to determine the basic spatiotemporal relationships between objects; Furthermore, statistical or sequence analysis is performed on the attribute information based on probabilistic rules to generate statistical spatiotemporal relationships; wherein, adjacent, disjoint, and merged relationships constitute a basic spatiotemporal relationship set; the basic relationships serve as the basis for inferring complex spatiotemporal relationships to expand and generate complex spatiotemporal relationship types, wherein the complex spatiotemporal relationship types include at least one of the following: covering relationship, inclusion relationship, co-position relationship, intersection / union relationship; the statistical spatiotemporal relationships include at least one of the following: evolutionary / sequence relationship and correlation relationship.

8. The data relationship generation method according to claim 7, characterized in that, The data relationship generation method also includes: relationship lifecycle management; When a new standardized data record is added to the database, perform the following operations: If the standardized data record is associated with an existing spatiotemporal relationship, the standardized data record is added as evidence to the evidence list of the spatiotemporal relationship, and the confidence score of the spatiotemporal relationship is recalculated based on the updated evidence list. Otherwise, based on the standardized data records and other existing records, the relationship is deduced through the deterministic or probabilistic rules. If a spatiotemporal relationship that does not yet exist in the data relationship network is deduced, a corresponding new spatiotemporal relationship data item is generated, and an initial evidence list and confidence score are created. Based on preset failure rules, the spatiotemporal relationships in the data relationship network are periodically checked, and spatiotemporal relationships with confidence scores below preset thresholds or without new evidence within preset time periods are marked as failures or archived. Preserve the complete chain of evidence and data source information for all spatiotemporal relationships to support historical retrospection and human intervention correction of the relationship derivation process.

9. A spatiotemporal fusion coding system, characterized in that, include: The data acquisition module is used to acquire spatial information, temporal information, and object identification information of multi-source heterogeneous target data objects; The spatial coding module is used to map the spatial information into spatial codes based on a geospatial grid coding system; wherein the geospatial grid coding system forms a multi-level grid by recursively dividing the Earth's space, and each grid has a unique coding identifier; The time encoding module is used to standardize the time information and map the standardized time information into time codes based on the time hierarchical encoding system. The object identifier generation module is used to inherit the attributes of the object identifier information and convert them into standardized attribute information based on the object encoding engine, so as to generate a unique object identifier for the target data object; The fusion encoding generation module is used to establish an association between the spatial encoding, the temporal encoding and the object identifier based on a preset bit-level fusion rule, and generate a spatiotemporal fusion encoding of the target data object, wherein the spatiotemporal fusion encoding is a structured binary encoding string that uniquely identifies the target data object; The data record generation module is used to bind the spatiotemporal fusion encoding with the attribute information of the target data object based on the standardized attribute information to obtain standardized data records, so as to construct a cross-system spatiotemporal relationship network.

10. A data relationship generation system, characterized in that, include: A data acquisition module is used to acquire standardized data records of multiple target data objects, wherein each of the standardized data records is generated based on the spatiotemporal fusion coding method described in any one of claims 1-4, and the standardized data record includes at least: spatiotemporal fusion coding and attribute information bound to the spatiotemporal fusion coding; The data partitioning module is used to partition the standardized data records according to a preset level based on the spatiotemporal fusion coding. The relationship deduction module is used to decode and / or perform logical operations on the spatiotemporal fusion encoding within each hierarchical partition, and deduce the spatiotemporal relationship between the attribute information of different data objects based on the results of the decoding and / or logical operation. A data relationship network construction module is used to construct a data relationship network based on the spatiotemporal relationship. Each spatiotemporal relationship data item in the data relationship network is associated with an evidence list and a confidence score. The evidence list records one or more standardized data records on which the spatiotemporal relationship is derived. The confidence score is obtained based on a preset scoring rule, which includes at least one of the following: the quality score of the data record in the evidence list, the strictness of the rule matching, the time decay factor, and the statistical score.

Images