A mineral resource intelligent management method and system based on multi-source data fusion

Abstract

This invention relates to the field of mining big data management technology, specifically to an intelligent management method and system for mineral resources based on multi-source data fusion. The method includes the following steps: extracting mineral resource business data from a multi-source mineral resource business database and writing it into source metadata; preprocessing and entity modeling the mineral resource business data to generate a unified mineral resource data model; aligning entities based on object recognition fields and constructing a mineral resource knowledge graph; parsing natural language management instructions through a supervisory agent; scheduling sub-agents to access the knowledge graph or corresponding database according to task type and data requirements; and fusing the results of sub-tasks to generate mineral resource management results. This invention, through unified modeling, entity alignment, graph association, and agent scheduling, enables cross-database tasks to be executed according to task type and data requirements, and adds traceability and verification information, giving the mineral resource management results a traceable and verifiable output structure.

Description

Technical Field

[0001] This invention relates to the field of mining big data management technology, and in particular to an intelligent management method and system for mineral resources based on multi-source data fusion. Background Technology

[0002] The field of big data management technology in the mining industry involves the collection, storage, integration, analysis, and management of mineral resource exploration data, reserve data, mining data, monitoring data, and related business data. It is primarily used for the unified organization and management of information throughout the entire lifecycle of mineral resources and is widely applied in industries such as natural resource management, mine development, geological exploration, resource supervision, and smart mine construction. Traditional intelligent management methods for mineral resources refer to the storage, querying, statistics, updating, and management of basic mineral resource information, exploration results information, production operation information, and resource supervision information through database systems, resource management platforms, geographic information systems, and information management systems to support information organization and business management in the mineral resource management process.

[0003] Traditional mineral resource management relies on business databases such as the national mineral resource information database, reserve database, rights confirmation database, and asset inventory database. The field standards, update cycles, and interface forms of each database are different. Managers need to query, compare, and summarize across systems during the process of mineral rights approval, reserve verification, and asset accounting. It is difficult to retain the data source relationship and verification status synchronously. The cross-database correlation results lack unified object relationship support, and the process of forming management conclusions lacks a continuous data link, which affects the consistency and traceability of the comprehensive mineral resource management results. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent management method and system for mineral resources based on multi-source data fusion. This system is used to uniformly extract, label, preprocess, model, align, and construct knowledge graphs from business data in multi-source business databases of mineral resources. Through the collaborative processing of the supervisory agent and sub-agents, it realizes the parsing of natural language management instructions, task scheduling, and result fusion. This solves the problems of unclear relationships between mineral resource business data in multi-database management scenarios, reliance on manual comparison for cross-database processing, and inconvenience in tracing the source and verification process of results.

[0005] To achieve the above objectives, the present invention adopts the following technical solution.

[0006] This invention provides an intelligent management method for mineral resources based on multi-source data fusion, the method comprising:

[0007] Extract mineral resource business data from the multi-source mineral resource business database and write source metadata into the mineral resource business data;

[0008] The mineral resource business data is preprocessed and entity modeled to generate a unified mineral resource data model, which represents the relationship between mineral resource management objects.

[0009] Based on the object identification field in the unified mineral resource data model, entity alignment is performed on the source records formed by the mineral resource business data, and a mineral resource knowledge graph is constructed based on the entity alignment results.

[0010] The supervisory agent parses natural language management instructions, determines the task decomposition results, and schedules matching sub-agents to access the corresponding database in the mineral resource knowledge graph or the mineral resource multi-source business database according to the task type and data requirements in the task decomposition results, so as to obtain the sub-task results.

[0011] The supervisory agent, based on the entity alignment relationship and the task constraints in the task decomposition results, fuses the sub-task results to generate mineral resource management results with attached result traceability information and verification information.

[0012] Furthermore, the extraction process of the multi-source business database of mineral resources includes: determining the data table range, extraction fields and field sources of each business database according to a preset database interface mapping table; extracting the mineral resource business data from each business database according to the data table range and the extraction fields; writing the database identifier, business table identifier, extraction time and record version into the source metadata, and binding the source metadata with the corresponding mineral resource business data.

[0013] Furthermore, before binding the source metadata to the corresponding mineral resource business data, a source integrity verification is performed on the mineral resource business data. The source integrity verification includes: verifying whether the database identifier matches the business table identifier; verifying whether the extracted field belongs to the field range recorded in the database interface mapping table; when the database identifier matches the business table identifier and the extracted field belongs to the field range, the source metadata is written to the corresponding mineral resource business data; when any condition is not met, the corresponding record is marked as a record to be reviewed.

[0014] Furthermore, the mineral resource business data that has passed the source integrity verification enters the preprocessing process, which includes: determining the standard field name and standard field type corresponding to the mineral resource business data according to the database interface mapping table; writing the original field values ​​in the mineral resource business data into the standard field corresponding to the standard field name; generating a standardized business record when the original field value matches the standard field type, and marking the corresponding field as a field to be converted when the original field value does not match the standard field type.

[0015] Furthermore, the entity modeling process includes: determining entity categories based on the business object types in the standardized business records; writing the object name field, ownership identifier field, spatial range field, and business time field from the standardized business records into the corresponding entity categories; and generating the unified mineral resource data model based on the subordinate relationships and business associations between the entity categories.

[0016] Furthermore, the object identification field includes the object name field, the ownership identifier field, the spatial range field, and the business time field. The process of entity alignment based on the object identification field includes: initially grouping the source records according to the object name field; comparing the ownership identifier field and the spatial range field within the same initial group; when the ownership identifier field is consistent and the spatial range field meets the overlap condition, the corresponding source record is determined as an aligned entity record of the same mineral resource management object; when the ownership identifier field is inconsistent or the spatial range field does not meet the overlap condition, the corresponding source record is retained as a source record to be confirmed.

[0017] Furthermore, the process of constructing the mineral resource knowledge graph based on the entity alignment results includes: generating mineral resource management object nodes based on the aligned entity records; determining the business association relationships between the mineral resource management object nodes based on the ownership identifier field and the business time field in the object identification field; and writing the mineral resource management object nodes and the business association relationships into a graph database to form the mineral resource knowledge graph accessed by the sub-agent.

[0018] Furthermore, the process by which the supervisory agent parses the natural language management instructions and schedules the matching sub-agents includes: extracting the management object, business action, time range, and output requirements from the natural language management instructions; determining the task type based on the business action; determining the data requirement based on the management object and the time range; when the data requirement points to an established entity relationship, scheduling the matching sub-agents to access the mineral resource knowledge graph; and when the data requirement points to a source record field, scheduling the matching sub-agents to access the corresponding database in the multi-source business database of mineral resources to obtain the sub-task result.

[0019] Furthermore, the process of merging the results of the subtasks includes: merging the results of the subtasks according to the management object and the time range to form candidate fusion results; matching the data source in the candidate fusion results with the source metadata, and generating the result traceability information based on the matching result; verifying the field integrity and source consistency in the candidate fusion results according to the task type, generating the verification information, and writing the result traceability information and the verification information into the mineral resource management result.

[0020] The present invention also provides an intelligent management system for mineral resources based on multi-source data fusion. The system is used to implement the above-mentioned intelligent management method for mineral resources based on multi-source data fusion. The system includes a data extraction module, a data modeling module, a graph construction module, an agent scheduling module, and a result fusion module.

[0021] The data extraction module extracts mineral resource business data from the multi-source mineral resource business database and writes source metadata into the mineral resource business data;

[0022] The data modeling module preprocesses and models the mineral resource business data to generate a unified mineral resource data model, which represents the relationships between mineral resource management objects.

[0023] The knowledge graph construction module performs entity alignment on the source records formed by the mineral resource business data based on the object identification field in the unified mineral resource data model, and constructs a mineral resource knowledge graph based on the entity alignment results.

[0024] The agent scheduling module parses natural language management instructions through the supervisor agent, determines the task decomposition results, and schedules the matched sub-agents to access the corresponding database in the mineral resource knowledge graph or the mineral resource multi-source business database according to the task type and data requirements in the task decomposition results, so as to obtain the sub-task results.

[0025] The result fusion module, through the supervisory agent, fuses the sub-task results according to the entity alignment relationship and the task constraints in the task decomposition results, and generates mineral resource management results with attached result traceability information and verification information.

[0026] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0027] By extracting mineral resource business data from a multi-source mineral resource business database and writing it into source metadata, subsequent processing can retain the data source chain; by forming a unified mineral resource data model through preprocessing and entity modeling, the relationships between mineral resource management objects have a unified expression basis; by entity alignment and mineral resource knowledge graph construction, records from different sources can establish associations around the same management object; by parsing natural language management instructions according to the supervisory agent and scheduling sub-agents, cross-database tasks are executed according to task type and data requirements; based on the results of fused sub-tasks and adding traceability and verification information, the mineral resource management results have an output structure that is traceable in origin and verifiable. Attached Figure Description

[0028] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 This is the main flowchart of the intelligent management of mineral resources in this invention;

[0030] Figure 2 This is a flowchart illustrating the data extraction and source metadata binding process of this invention.

[0031] Figure 3 This invention provides a flowchart for aligning the unified mineral resource data model with entities.

[0032] Figure 4 This is a flowchart illustrating the construction process of the mineral resource knowledge graph in this invention.

[0033] Figure 5 This is a flowchart of the intelligent agent scheduling and result fusion process of the present invention;

[0034] Figure 6 This is a schematic diagram of the system of the present invention. Detailed Implementation

[0035] The embodiments of the present invention are described below with reference to the accompanying drawings. Subsequent embodiments use the following unified terminology: multi-source business database of mineral resources, mineral resource business data, source metadata, unified mineral resource data model, object identification field, source record, entity alignment result, mineral resource knowledge graph, supervisor agent, sub-agent, task decomposition result, sub-task result, result traceability information, verification information, and mineral resource management result. The multi-source business database of mineral resources represents multiple business databases related to mineral resource management; mineral resource business data represents data content extracted from the multi-source business database of mineral resources and participating in subsequent processing; source metadata represents information recording data sources, extraction status, and version status; the unified mineral resource data model represents the data organization structure representing mineral resource management objects and the relationships between them; and the mineral resource knowledge graph represents object nodes and business relationships formed by entity alignment results. The step numbers in subsequent embodiments are only used to distinguish processing flows and do not constitute numerical limitations other than the execution order. The figure numbers in the accompanying drawings are only used to illustrate the correspondence between the figures and do not constitute a limitation on the technical scope. This embodiment does not impose specific limitations on the deployment location of each business database, the type of database product, the implementation method of the graph database, the type of natural language parsing model, or the number of sub-agents. The above content can be determined within the scope of the original disclosure based on the actual deployment conditions of the mineral resource management business system.

[0036] Please see Figures 1 to 5 This embodiment provides an intelligent management method for mineral resources based on multi-source data fusion. This method is applied to the cross-database fusion, object association, graph construction, natural language task parsing, and management result output of mineral resource business data.

[0037] A multi-source business database for mineral resources can include a national database of mineral resources, a database of mineral resources reserves, a database of mineral resources ownership registration, and a database of mineral resources asset inventory. It can also be other mineral resource management business databases formed within the same business boundary. This method unifies the extraction, source binding, standardization, entity modeling, entity alignment, and graph-based association of data from different sources, with different field definitions, and different ways of expressing business objects. This enables the supervisory agent to schedule sub-agents to access the mineral resource knowledge graph or corresponding database based on natural language management instructions, and to generate mineral resource management results with accompanying traceability and verification information based on the results of the sub-tasks.

[0038] S1: Extract mineral resource business data from the multi-source mineral resource business database and write source metadata for the mineral resource business data.

[0039] In S1, the multi-source business database of mineral resources serves as the input object, and the mineral resource business data serves as the extraction object. Before performing extraction, a preset database interface mapping table is read. This table records the range of data tables that can be extracted from each business database, the fields to be extracted, and the source of those fields. The range of data tables limits the business tables participating in the extraction, the fields to be extracted limit the content of the fields obtained from the business tables, and the source of the fields indicates the original field location and business meaning of the extracted fields in the corresponding business database. The preset database interface mapping table avoids data object confusion caused by inconsistencies in field names between different business databases.

[0040] The data table range, extracted fields, and field sources for each business database are determined according to the preset database interface mapping table.

[0041] The pre-defined database interface mapping table can set up field mapping relationships for the National Mineral Resources Information Database, the Mineral Resources Reserves Database, the Mineral Resources Ownership Registration Database, and the Mineral Resources Asset Inventory Database, respectively. Fields related to mining area identification, mineral type, and resource status in the National Mineral Resources Information Database; fields related to reserve year, recoverable reserves, and cumulative extraction status in the Mineral Resources Reserves Database; fields related to licenses, mining right holders, validity periods, and mining area boundaries in the Mineral Resources Ownership Registration Database; and fields related to assessment reports, corresponding licenses, assessment benchmark status, and net asset value status in the Mineral Resources Asset Inventory Database can all be used as components of the field source. These fields are not limited to fixed names; the actual names of the fields are determined by the field definitions in the corresponding business databases, and the meanings of the fields are uniformly explained by the pre-defined database interface mapping table.

[0042] Mineral resource business data is extracted from each business database based on the data table range and the fields to be extracted.

[0043] The extraction process uses the data table range as the extraction boundary and the extracted fields as the data content boundary. For data records that belong to the data table range and contain the extracted fields, the system extracts the corresponding field content as mineral resource business data. For data tables that do not belong to the data table range, or fields that do not belong to the extracted field range, the system does not include the corresponding content in this mineral resource business data. Mineral resource business data can retain its original field values, or it can enter the subsequent preprocessing process after extraction, with the original field values ​​used for subsequent standard field conversion and source tracing.

[0044] Write the database identifier, business table identifier, extraction time, and record version into the source metadata, and bind the source metadata with the corresponding mineral resource business data.

[0045] Database identifiers distinguish the database source of mineral resource business data, business table identifiers distinguish the business table source of the mineral resource business data, extraction time characterizes the time state when the mineral resource business data enters the processing chain of this method, and record version characterizes the version state of the same business record at the time of extraction. After the source metadata is bound to the mineral resource business data, subsequent preprocessing, entity modeling, entity alignment, graph construction, sub-agent access, and result fusion processes can all use the source metadata to form a traceability chain. The source metadata does not change the business meaning of the mineral resource business data, but serves as an auxiliary data structure for subsequent verification and traceability.

[0046] Before binding the source metadata to the corresponding mineral resource business data, a source integrity check is performed. The source integrity check is used to confirm the correspondence between the source fields of the mineral resource business data and the preset database interface mapping table.

[0047] Verify whether the database identifier matches the business table identifier.

[0048] The matching relationship between database identifiers and business table identifiers is derived from a pre-defined database interface mapping table. When a business table identifier belongs to the range of data tables corresponding to a database identifier, the database identifier and business table identifier meet the matching condition; when a business table identifier does not belong to the range of data tables corresponding to a database identifier, the database identifier and business table identifier do not meet the matching condition. This verification is used to prevent the incorrect merging of business tables with the same name in different business databases.

[0049] Verify whether the extracted fields fall within the range of fields recorded in the database interface mapping table.

[0050] The extracted fields are compared with the field range recorded in the database interface mapping table. If the extracted field falls within the field range of the corresponding business database and business table, the extracted field meets the field range condition; if the extracted field does not fall within the corresponding field range, the extracted field does not meet the field range condition. This verification ensures that the mineral resource business data entering subsequent processing has a clear field source and business meaning.

[0051] When the database identifier matches the business table identifier and the extracted field falls within the field range, the source metadata is written into the corresponding mineral resource business data. When any condition is not met, the corresponding record is marked as a record to be reviewed.

[0052] The "Pending Review Record" represents the output status of a failed source integrity check. This record does not directly enter the standardized business record generation process. It can be subsequently confirmed by administrators or existing business systems, or it can re-enter the source integrity check process after confirmation. This embodiment does not specifically limit the manual processing interface, processing time, or personnel permissions for the "Pending Review Record." The "Pending Review Record" is only used to illustrate the output status and data flow when the source integrity check fails.

[0053] S2: Preprocess and model mineral resource business data to generate a unified mineral resource data model, which represents the relationships between mineral resource management objects.

[0054] In S2, mineral resource business data that has passed source integrity verification enters the preprocessing process. The goal of the preprocessing process is to convert the raw field values, which are expressed differently in different business databases, into standardized business records that can be incorporated into the unified mineral resource data model. The goal of the entity modeling process is to determine the entity categories based on the business object types in the standardized business records and to write the fields that can identify mineral resource management objects and their relationships into the corresponding entity categories.

[0055] The standard field names and standard field types corresponding to the mineral resource business data are determined based on the database interface mapping table.

[0056] Standard field names are used to unify fields with the same business meaning but different names across different business databases. Standard field types are used to define the field status that raw field values ​​must meet when entering a standard field. Standard field types are not limited by specific numerical values, but are categorized based on the business meaning of the field, such as name fields, identifier fields, spatial range fields, business time fields, ownership fields, reserve status fields, and asset status fields. Standard field names and standard field types are provided by the database interface mapping table, which serves as the common basis for both the extraction and preprocessing processes.

[0057] Write the original field values ​​from the mineral resource business data into the standard field corresponding to the standard field name.

[0058] The original field values ​​serve as the specific field content of mineral resource business data. For a given original field value, the system determines the corresponding standard field name based on the database interface mapping table and writes the original field value into the standard field. This writing process preserves the binding relationship between the original field value and the source metadata, ensuring that the content in the standard field can still be traced back to the original business database, the original business table, and the record version at the time of extraction.

[0059] Standardized business records are generated when the original field value matches the standard field type; when the original field value does not match the standard field type, the corresponding field is marked as a field to be converted.

[0060] Standardized business records are used as input for entity modeling. Whether the original field values ​​match the standard field types can be determined based on field format, field semantics, field source, and business table identifier. When the original field values ​​can be entered into the corresponding standard field and maintain consistency with the field's business meaning, a standardized business record is generated; when the original field values ​​cannot directly meet the standard field type requirements, the corresponding field is marked as a field to be converted. Fields to be converted indicate that the field needs to undergo field definition confirmation or conversion processing.

[0061] The entity category is determined based on the business object type in the standardized business records.

[0062] Business object types are used to distinguish the mineral resource management objects described by standardized business records. Entity categories can correspond to business objects such as mining areas, mining rights, resource reserves, and asset values, or they can be set according to the entity categories defined in the unified mineral resource data model within the same business boundary. The business object type can be determined by the business table identifier, standard field names, and the business meaning of the fields. After the entity category is determined, the standardized business record enters the modeling process for the corresponding entity category.

[0063] Write the object name field, ownership identifier field, spatial range field, and business time field from the standardized business record into the corresponding entity category.

[0064] The object name field represents the name status of the mineral resource management object; the ownership identifier field represents the ownership status of the mineral resource management object; the spatial range field represents the spatial coverage status of the mineral resource management object; and the business time field represents the time status of the mineral resource management object in the business process. The object name, ownership identifier, spatial range, and business time fields together form an important basis for subsequent entity alignment. Different entity categories can accept different combinations of fields, but the fields used for entity alignment should be able to support object identification and association determination between source records.

[0065] A unified mineral resource data model is generated based on the hierarchical and business relationships between entity categories.

[0066] Subordination relationships are used to characterize the hierarchical affiliation between different entity categories, while business relationship relationships are used to characterize the business connections between different entity categories in the mineral resource management process. The unified mineral resource data model is not limited to a specific database structure, but rather is a data model used to uniformly express the relationships between mineral resource management objects. After generating the unified mineral resource data model, source records can enter the entity alignment process based on the unified entity category and object identification fields.

[0067] S3: Based on the object identification field in the unified mineral resource data model, perform entity alignment on the source records formed by mineral resource business data, and construct a mineral resource knowledge graph based on the entity alignment results.

[0068] In S3, a source record refers to a record formed from mineral resource business data after source integrity verification, preprocessing, and entity modeling, retaining the source binding relationship. Object identification fields include object name, ownership identifier, spatial range, and business time. Entity alignment is used to determine whether different source records correspond to the same mineral resource management object. The entity alignment result is used to generate object nodes and business relationships in the mineral resource knowledge graph.

[0069] Initially group the source records by the object name field.

[0070] The object name field serves as the initial grouping criterion. The system compares the object name field in the source records and groups source records whose object names match or have a corresponding relationship into the same initial group. This initial grouping narrows down the scope of subsequent comparisons of the ownership identifier and spatial range fields. For source records where the object name field cannot form a clear group, the system can retain the source record and await subsequent business review.

[0071] Within the same initial group, the ownership identifier field and the spatial range field are compared. When the ownership identifier field is consistent and the spatial range field meets the overlap condition, the corresponding source record is identified as the aligned entity record of the same mineral resource management object.

[0072] The ownership identifier field is used to determine whether the ownership entities involved in the source records are consistent, and the spatial range field is used to determine whether the spatial range involved in the source records can correspond to the same mineral resource management object. The overlap condition is used to indicate that there is a spatial correspondence between the spatial range fields that can support the determination of the same object. Only when the ownership identifier field is consistent and the spatial range field meets the overlap condition will the system determine the corresponding source records as aligned entity records of the same mineral resource management object. The aligned entity records serve as input for the construction of the mineral resource knowledge graph.

[0073] The overlap condition for spatial range fields means that, after unifying the spatial representation, the spatial range fields of source records within the same initial group can form a spatial correspondence with the same mineral resource management object. Unifying the spatial representation includes converting mining area coordinates, boundary descriptions, or range registration information into comparable spatial boundary information while preserving the binding relationship with the source metadata. When the spatial boundary information of two source records overlaps, contains, or is contained within each other, and this spatial correspondence does not conflict with the ownership identifier field or the business time field, the spatial range field is deemed to meet the overlap condition. When the spatial range field of any source record is missing, the spatial representation cannot be unified, the spatial boundary information cannot be parsed, or the spatial boundary information of two source records is separated, the spatial range field is deemed not to meet the overlap condition, and the corresponding source record is retained as a source record awaiting confirmation. The overlap condition is used to determine whether source records in different databases can correspond to the same mineral resource management object in terms of spatial range; complete consistency of spatial range is not a necessary condition.

[0074] When the ownership identifier field is inconsistent or the spatial range field does not meet the overlap condition, the corresponding source record will be retained as a source record to be confirmed.

[0075] A source record to be confirmed refers to a source record that cannot be directly identified as belonging to the same mineral resource management object through the current object identification field. Source records to be confirmed are not directly merged into aligned entity records to avoid incorrectly merging different mineral resource management objects. Source metadata can be retained for these records and used for review or field supplementation in subsequent management processes. This embodiment does not specify a particular method for reviewing source records to be confirmed.

[0076] Mineral resource management object nodes are generated based on the aligned entity records.

[0077] Mineral resource management object nodes are used to represent entity nodes in the mineral resource knowledge graph. The object name, ownership identifier, spatial range, and business time fields corresponding to the aligned entity records can serve as the attribute source for mineral resource management object nodes. Mineral resource management object nodes maintain a connection with the source metadata, enabling the business content corresponding to the node to be traced back to its extraction source.

[0078] The business relationships between mineral resource management object nodes are determined based on the ownership identifier field and the business time field in the object identification field.

[0079] The ownership identifier field is used to determine the ownership association status between mineral resource management object nodes, while the business time field is used to determine the time association status between them. Business associations can represent the subordination, generation, valuation, or other business relationships already defined in the unified mineral resource data model between entities such as mining areas, mining rights, resource reserves, and asset values. Business associations do not change the original business meaning of the source record; rather, they are used to construct association paths accessible to intelligent agents.

[0080] Mineral resource management object nodes and their business relationships are written into a graph database to form a mineral resource knowledge graph accessible by sub-intelligent agents.

[0081] Graph databases are used to store mineral resource management object nodes and business relationships. Once the mineral resource knowledge graph is formed, sub-agents can access established entity relationships through object nodes and business relationships, and can also trace back to the corresponding database in the multi-source business database of mineral resources through source metadata in node attributes. The specific product type and deployment form of the graph database are not specifically limited, as long as it can store mineral resource management object nodes, business relationships, and related traceability relationships.

[0082] S4: The supervisory agent parses the natural language management instructions, determines the task decomposition results, and schedules the matched sub-agents to access the corresponding databases in the mineral resource knowledge graph or the multi-source business database of mineral resources according to the task type and data requirements in the task decomposition results, so as to obtain the sub-task results.

[0083] In S4, natural language management instructions serve as input to the supervisory agent. The supervisory agent identifies the management object, business action, time frame, and output requirements from these instructions, thereby determining the task type and data needs. Sub-agents execute sub-tasks corresponding to specific business types, including national conditions analysis, reserve dynamics analysis, rights confirmation and compliance analysis, and asset accounting analysis.

[0084] Extract the management object, business action, time range, and output requirements from natural language management instructions.

[0085] The managed object represents the mining area, mining rights, resource reserves, or asset value targeted by the natural language management command. The business action represents the business processing intent such as querying, verifying, statistically analyzing, associating, generating reports, or generating early warnings. The time range represents the time conditions of the business involved in the natural language management command. The output requirements represent the fields, formats, or traceability status required for the mineral resource management results. After extracting the above content, the supervisor agent writes it into the basic fields of the task decomposition results.

[0086] Determine the task type based on business actions, and determine the data requirements based on the management objects and time range.

[0087] The task type determines which sub-agent will execute the corresponding sub-task, while the data requirement determines whether the sub-agent accesses the mineral resource knowledge graph or the corresponding database within the multi-source business database of mineral resources. When business actions involve object associations, ownership relationships, reserve relationships, or asset relationships, the task type can be used for relational queries or comprehensive analysis; when business actions involve source field verification or original record confirmation, the data requirement can point to the source record field. The task type and data requirement, as important components of the task decomposition results, provide a basis for subsequent scheduling.

[0088] When data requests point to established entity relationships, the matching sub-agents are scheduled to access the mineral resource knowledge graph.

[0089] Established entity relationships refer to the business associations between mineral resource management object nodes that have been written into the mineral resource knowledge graph. When data requirements can be obtained through object nodes and business associations in the mineral resource knowledge graph, the supervisor agent schedules matching sub-agents to access the mineral resource knowledge graph. The sub-agents retrieve the corresponding object nodes, business associations, and node attributes in the mineral resource knowledge graph based on the task type.

[0090] When the data requirement points to the source record field, the scheduled matching sub-agent accesses the corresponding database in the multi-source business database of mineral resources to obtain the sub-task result.

[0091] The source record field refers to the original source field that needs to be verified in the corresponding database within the multi-source business database of mineral resources. When data requirements cannot be fulfilled solely through established entity relationships, or when the task requires verification of source fields, the supervisor agent schedules a matching sub-agent to access the corresponding database. The corresponding database is determined by the database identifier and business table identifier in the source metadata. The sub-agent reads the field content that matches the task type and data requirements and returns the sub-task result. The sub-task result retains the correspondence between the task source, the accessed object, and the data source for subsequent fusion.

[0092] S5: Through the supervisory agent, based on the entity alignment relationship and the task constraints in the task decomposition results, the sub-task results are integrated to generate mineral resource management results with attached result traceability information and verification information.

[0093] In S5, subtask results serve as fusion input, entity alignment relationships are used as the basis for merging objects under the same mineral resource management system, and task constraints in the task decomposition results are used as the basis for filtering and verification. Task constraints can be derived from management objects, business actions, timeframes, and output requirements extracted from natural language management instructions. The fusion process generates candidate fusion results, result traceability information, verification information, and mineral resource management results.

[0094] The results of subtasks are merged according to the management object and time range to form candidate fusion results.

[0095] The management object is used to determine the mineral resource management object to which the subtask result belongs, and the time range is used to determine whether the subtask result meets the time conditions specified by the natural language management instructions. The supervisor agent merges the subtask results belonging to the same management object and meeting the time range requirements to form candidate fusion results. The candidate fusion results are intermediate results for traceability matching and verification processing and are not directly output as the final mineral resource management result.

[0096] The data sources in the candidate fusion results are matched with the source metadata, and result traceability information is generated based on the matching results.

[0097] The data sources in the candidate fusion results include source information generated when sub-agents access the mineral resource knowledge graph or corresponding database. The supervisor agent matches the data source with the database identifier, business table identifier, extraction time, and record version in the source metadata. When the match is successful, result traceability information that characterizes the data source chain is generated; when the match is unsuccessful, the corresponding content in the candidate fusion results can enter a pending review state. The result traceability information is used to explain the source relationships of each field or object in the mineral resource management results.

[0098] The integrity and source consistency of the fields in the candidate fusion results are verified according to the task type, and verification information is generated.

[0099] Field completeness indicates whether the candidate fusion result contains the necessary fields required by the task type, while source consistency indicates whether different data sources in the candidate fusion result are consistent with the same management object, the same task type, and the same time range. The supervisory agent determines the set of fields to be verified and the source consistency requirements according to the task type. When verification passes, verification information indicating a passed status is generated; when verification fails, verification information indicating a pending review or incomplete field status is generated.

[0100] The traceability and verification information of the results will be written into the mineral resource management results.

[0101] The mineral resource management results, as the final output, include integrated business content, along with result traceability and verification information. The result traceability information is used for subsequent management processes to verify data sources, while the verification information indicates the completeness and consistency of fields within the mineral resource management results. The mineral resource management results can be used for comprehensive queries, business verification, report generation, or early warning output.

[0102] Please see Figure 6 A mineral resource intelligent management system based on multi-source data fusion is proposed. This system is used to implement the aforementioned intelligent management method for mineral resources based on multi-source data fusion. The system includes a data extraction module, a data modeling module, a graph construction module, an agent scheduling module, and a result fusion module.

[0103] The data extraction module is used to extract mineral resource business data from a multi-source mineral resource business database and write source metadata for the data. The module receives business data input from the database, reads a preset database interface mapping table to determine the data table range, extracted fields, and field sources, and outputs the mineral resource business data bound to the source metadata. The module can also perform source integrity verification, outputting verified mineral resource business data to the data modeling module and marking records that fail verification as pending review.

[0104] The data modeling module is used to preprocess and model entities in mineral resource business data, generating a unified mineral resource data model. This unified model represents the relationships between mineral resource management objects. The data modeling module receives mineral resource business data output from the data extraction module, determines standard field names and types based on the database interface mapping table, writes the original field values ​​into the corresponding standard fields, and generates standardized business records. The data modeling module also determines entity categories based on the business object types in the standardized business records, writes the object name field, ownership identifier field, spatial range field, and business time field into the corresponding entity categories, and outputs the unified mineral resource data model.

[0105] The graph construction module unifies the object identification fields in the mineral resource data model, aligns source records formed from mineral resource business data into entities, and constructs a mineral resource knowledge graph based on the alignment results. The module receives the unified mineral resource data model and source records output by the data modeling module, performs initial grouping according to the object name field, compares the ownership identifier field and spatial range field within the same initial group, and generates aligned entity records or source records to be confirmed. Based on the aligned entity records, the module generates mineral resource management object nodes, determines business relationships based on the ownership identifier field and business time field, and writes the mineral resource management object nodes and business relationships into the graph database, forming the mineral resource knowledge graph.

[0106] The agent scheduling module is used by the supervisor agent to parse natural language management instructions, determine the task decomposition results, and schedule matching sub-agents to access the corresponding databases in the mineral resource knowledge graph or the multi-source business database of mineral resources according to the task type and data requirements in the task decomposition results, thereby obtaining the sub-task results. The agent scheduling module receives natural language management instructions, extracts the management object, business action, time range, and output requirements from the instructions, and determines the task type and data requirements. When the data requirement points to an established entity relationship, the agent scheduling module schedules the matching sub-agents to access the mineral resource knowledge graph; when the data requirement points to a source record field, the agent scheduling module schedules the matching sub-agents to access the corresponding database in the multi-source business database of mineral resources.

[0107] The result fusion module, through the supervisory agent, fuses sub-task results according to entity alignment relationships and task constraints in the task decomposition results, generating mineral resource management results with accompanying result traceability and verification information. The result fusion module receives sub-task results output by the agent scheduling module, merges them according to management objects and time ranges to form candidate fusion results; matches the data sources in the candidate fusion results with source metadata to generate result traceability information; verifies the field integrity and source consistency in the candidate fusion results according to the task type, generating verification information; and writes the result traceability and verification information into the mineral resource management results.

[0108] The specific processes, field states, processing order, judgment conditions, module collaboration relationships, and carrier inheritance relationships described in the above embodiments are used to explain the possible implementations of the present invention and should not limit the present invention to the specific embodiments listed. Without departing from the scope of the claims and the original disclosure, any equivalent substitutions, equivalent modifications, equivalent combinations, order adjustments, corresponding module replacements, equivalent changes in field names, equivalent inheritance of the executing entity, or equivalent changes in the carrier form that can be conceived by those skilled in the art should fall within the protection scope of this application; however, it should not be extended to unclaimed topics, nor should it alter the substantive correspondence between the multi-source business database of mineral resources, mineral resource business data, source metadata, unified mineral resource data model, mineral resource knowledge graph, supervisory agent, sub-agent, sub-task results, and mineral resource management results through name changes.

Claims

1. A method for intelligent management of mineral resources based on multi-source data fusion, characterized in that, The method includes: Extract mineral resource business data from the multi-source mineral resource business database and write source metadata into the mineral resource business data; The mineral resource business data is preprocessed and entity modeled to generate a unified mineral resource data model, which represents the relationship between mineral resource management objects. Based on the object identification field in the unified mineral resource data model, entity alignment is performed on the source records formed by the mineral resource business data, and a mineral resource knowledge graph is constructed based on the entity alignment results. The supervisory agent parses natural language management instructions, determines the task decomposition results, and schedules matching sub-agents to access the corresponding database in the mineral resource knowledge graph or the mineral resource multi-source business database according to the task type and data requirements in the task decomposition results, so as to obtain the sub-task results. The supervisory agent, based on the entity alignment relationship and the task constraints in the task decomposition results, fuses the sub-task results to generate mineral resource management results with attached result traceability information and verification information.

2. The intelligent management method for mineral resources based on multi-source data fusion according to claim 1, characterized in that, The extraction process of the multi-source business database of mineral resources includes: The data table range, extracted fields, and field sources for each business database are determined according to the preset database interface mapping table. Mineral resource business data is extracted from each business database according to the data table range and the extraction fields. Write the database identifier, business table identifier, extraction time, and record version into the source metadata, and bind the source metadata with the corresponding mineral resource business data.

3. The intelligent management method for mineral resources based on multi-source data fusion according to claim 2, characterized in that, Before the source metadata is bound to the corresponding mineral resource business data, the source integrity of the mineral resource business data is verified. The source integrity verification includes: Verify whether the database identifier matches the business table identifier; Verify whether the extracted fields belong to the field range recorded in the database interface mapping table; When the database identifier matches the business table identifier and the extracted field belongs to the field range, the source metadata is written into the corresponding mineral resource business data. When any condition is not met, the corresponding record is marked as a record to be reviewed.

4. The intelligent management method for mineral resources based on multi-source data fusion according to claim 3, characterized in that, The mineral resource business data that has passed the source integrity verification enters the preprocessing process, which includes: The standard field names and standard field types corresponding to the mineral resource business data are determined based on the database interface mapping table. Write the original field values ​​from the mineral resource business data into the standard field corresponding to the standard field name; A standardized business record is generated when the original field value matches the standard field type; when the original field value does not match the standard field type, the corresponding field is marked as a field to be converted.

5. The intelligent management method for mineral resources based on multi-source data fusion according to claim 4, characterized in that, The entity modeling process includes: The entity category is determined based on the business object type in the standardized business record; Write the object name field, ownership identifier field, spatial range field, and business time field from the standardized business record into the corresponding entity category; The unified mineral resource data model is generated based on the hierarchical and business relationships between the entity categories.

6. The intelligent management method for mineral resources based on multi-source data fusion according to claim 5, characterized in that, The object identification field includes the object name field, the ownership identifier field, the spatial range field, and the business time field. The process of entity alignment based on the object identification field includes: The source records are initially grouped according to the object name field; Within the same initial group, the ownership identifier field and the spatial range field are compared. When the ownership identifier field is consistent and the spatial range field meets the overlap condition, the corresponding source record is determined as an aligned entity record of the same mineral resource management object. When the ownership identifier field is inconsistent or the spatial range field does not meet the overlap condition, the corresponding source record will be retained as a source record to be confirmed.

7. The intelligent management method for mineral resources based on multi-source data fusion according to claim 6, characterized in that, The process of constructing the mineral resource knowledge graph based on the entity alignment results includes: Mineral resource management object nodes are generated based on the aligned entity records; The business association relationship between the mineral resource management object nodes is determined based on the ownership identifier field and the business time field in the object identification field; The mineral resource management object nodes and their business relationships are written into a graph database to form the mineral resource knowledge graph accessed by the sub-agent.

8. The intelligent management method for mineral resources based on multi-source data fusion according to claim 1, characterized in that, The process by which the supervisory agent parses the natural language management instructions and schedules the matching sub-agents includes: Extract the management object, business action, time range, and output requirements from the natural language management instructions; The task type is determined based on the business action, and the data requirements are determined based on the management object and the time range; When the data request points to an established entity relationship, the matched sub-agent is scheduled to access the mineral resource knowledge graph. When the data request points to a source record field, the matched sub-agent is scheduled to access the corresponding database in the multi-source business database of mineral resources to obtain the sub-task result.

9. The intelligent management method for mineral resources based on multi-source data fusion according to claim 8, characterized in that, The process of fusing the results of the subtasks includes: The results of the subtasks are merged according to the management object and the time range to form candidate fusion results; The data sources in the candidate fusion results are matched with the source metadata, and the result traceability information is generated based on the matching results; The integrity and source consistency of the fields in the candidate fusion results are verified according to the task type, the verification information is generated, and the result traceability information and the verification information are written into the mineral resource management results.

10. A mineral resource intelligent management system based on multi-source data fusion, the system being used to implement the mineral resource intelligent management method based on multi-source data fusion as described in any one of claims 1-9, characterized in that, The system includes: The data extraction module extracts mineral resource business data from the multi-source mineral resource business database and writes source metadata for the mineral resource business data; The data modeling module preprocesses and models the mineral resource business data to generate a unified mineral resource data model, which represents the relationships between mineral resource management objects. The knowledge graph construction module performs entity alignment on the source records formed by the mineral resource business data based on the object identification field in the unified mineral resource data model, and constructs a mineral resource knowledge graph based on the entity alignment results. The agent scheduling module parses natural language management instructions through the supervisor agent, determines the task decomposition results, and schedules the matched sub-agents to access the corresponding database in the mineral resource knowledge graph or the mineral resource multi-source business database according to the task type and data requirements in the task decomposition results, so as to obtain the sub-task results. The result fusion module, through the supervisory agent, fuses the sub-task results according to the entity alignment relationship and the task constraints in the task decomposition results, and generates mineral resource management results with attached result traceability information and verification information.

Images