A method and system for remote sensing big data governance supporting heterogeneous storage

By establishing logical data objects and consumption relationships, remote sensing data governance units are formed. The storage location is adjusted according to the capacity status of heterogeneous storage sources, which solves the problem of low data retrieval efficiency in heterogeneous storage environments and realizes dynamic storage adjustment and efficient data access.

CN123115282APending Publication Date: 2026-11-17SUZHOU ZHONGKE IMAGE SKY REMOTE SENSING TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611414858.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-12
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing remote sensing big data governance methods are ill-suited to adapting to changes in actual consumption relationships in production processes within heterogeneous storage environments, resulting in low data retrieval efficiency and difficulty in dynamically adjusting data organization.

Method used

By establishing logical data objects and consumption relationships, determining the data governance coupling edge, forming a remote sensing data governance unit, and determining the target aggregation storage source based on the capacity status of heterogeneous storage sources, dynamically adjusting the data storage location, and forming a temporary collaborative copy to adapt to changes in production processes.

Benefits of technology

It enables dynamic storage and adjustment of remote sensing data, ensuring that data retrieval is consistent with the actual relationship of production processes, thereby improving data access efficiency and adaptability.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123115282A_ABST
    Figure CN123115282A_ABST
Patent Text Reader

Abstract

This invention belongs to the field of industrial data processing technology and discloses a method and system for remote sensing big data governance that supports heterogeneous storage. The method includes: establishing logical data objects and consumption relationships; determining the set of remaining consumption processes, dual coupling quantities, and data governance coupling edges based on the execution status; forming candidate connected components and splitting them into remote sensing data governance units; determining the target aggregation storage source; replicating to form a temporary collaborative copy and updating the set of effective physical storage locations; updating the set of remaining consumption processes after the process is completed, establishing new logical data objects, reconstructing the remote sensing data governance units, re-determining the target aggregation storage source, and updating the temporary collaborative copy. This invention achieves dynamic collaborative governance of the remote sensing production data storage status and the actual process consumption relationship.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of industrial data processing technology, specifically relating to a remote sensing big data governance method and system that supports heterogeneous storage. Background Technology

[0002] As the scale of remote sensing production tasks continues to expand, the amount of raw remote sensing images, control data, and intermediate data generated during industrial data processing is constantly increasing. This results in large data volumes, long processing chains, and continuous data access relationships between different processes. Actual production systems typically connect to different types of storage resources simultaneously, such as network-attached storage, object storage, and distributed storage. These storage resources differ in access protocols, storage capacity, and physical location, and data from the same production task may be distributed across multiple storage sources.

[0003] Existing remote sensing big data governance methods often focus on unified access, directory management, or fixed storage strategies, neglecting the actual consumption relationships of data during the production process. For data with spatially related processing scope that still needs to be accessed by the same subsequent processes, continued distributed storage requires access across multiple storage locations when preparing input data. Conversely, centralized storage according to fixed datasets or directories struggles to adapt to the constantly changing data access relationships after production processes are completed. Furthermore, new intermediate data is continuously generated during production, altering the relationships between existing data. Summary of the Invention

[0004] This invention provides a remote sensing big data governance method and system that supports heterogeneous storage, solving the technical problem that the fixed storage organization of remote sensing production data in a heterogeneous storage environment is difficult to dynamically adjust with changes in the actual consumption relationship of production processes.

[0005] This invention provides a remote sensing big data governance method that supports heterogeneous storage from different sources, comprising the following steps: Acquire heterogeneous storage sources, remote sensing production data, production processes and process input-output relationships, establish logical data objects that include spatial processing range, data capacity and effective physical storage location set, and establish consumption relationships based on process input-output relationships; Obtain the execution status of the production process, determine the set of remaining consumption processes based on the consumption relationship and the execution status, and determine the dual coupling quantity based on the spatial processing range of any two logical data objects and the set of remaining consumption processes, and establish a data governance coupling edge; Candidate connected components are formed based on the data governance coupling edges, and the candidate connected components are split according to the common remaining consumption process to obtain a set of remote sensing data governance units. Based on the available storage capacity of heterogeneous storage sources, the data capacity of logical data objects, and the set of effective physical storage locations, the target aggregation storage source for each remote sensing data governance unit is determined. Logical data objects not located in the target aggregation storage source are copied to the target aggregation storage source to form temporary collaborative copies, and the set of valid physical storage locations is updated. After the production process is completed, the set of remaining consumption processes is updated, new logical data objects are created for the intermediate data generated by the production process, and the data governance coupling edge and remote sensing data governance unit are redefined. Based on the redefined remote sensing data governance unit, the target aggregation storage source is redefined, and logical data objects not located in the target aggregation storage source are copied to the target aggregation storage source to form a new temporary collaborative copy. The original temporary collaborative copy is then converged, and the set of effective physical storage locations is updated.

[0006] This invention also provides a remote sensing big data governance system that supports heterogeneous storage from different sources, comprising: The basic relationship construction module is used to acquire heterogeneous storage sources, remote sensing production data, production processes and process input-output relationships, establish logical data objects containing spatial processing range, data capacity and effective physical storage location set, and establish consumption relationships based on process input-output relationships; The coupling relationship determination module is used to obtain the execution status of the production process, determine the set of remaining consumption processes based on the consumption relationship and the execution status, and determine the dual coupling quantity based on the spatial processing range of any two logical data objects and the set of remaining consumption processes, and establish a data governance coupling edge. The governance unit construction module is used to form candidate connected components based on the data governance coupling edges, and to split the candidate connected components according to the common remaining consumption process to obtain a set of remote sensing data governance units. The aggregation storage determination module is used to determine the target aggregation storage source for each remote sensing data governance unit based on the available storage capacity of heterogeneous storage sources, the data capacity of logical data objects, and the set of effective physical storage locations. The collaborative copy formation module is used to copy logical data objects that are not located in the target aggregation storage source to the target aggregation storage source to form temporary collaborative copies, and update the set of valid physical storage locations; The closed-loop reconstruction module is used to update the set of remaining consumption processes after the production process is completed, establish new logical data objects for the intermediate data generated by the production process, and redetermine the data governance coupling edge and remote sensing data governance unit; redetermine the target aggregation storage source according to the reconstructed remote sensing data governance unit, copy the logical data objects not located in the target aggregation storage source to the target aggregation storage source to form a new temporary collaborative copy, converge the original temporary collaborative copy, and update the set of effective physical storage locations.

[0007] The beneficial effects of this invention are as follows: This invention establishes logical data objects based on the spatial processing range of remote sensing production data and the actual consumption relationship of production processes, and forms a set of remaining consumption processes in conjunction with the execution status of production processes, enabling data governance relationships to change with the actual production progress. By determining the data governance coupling relationship through spatial association and remaining consumption relationships, logical data objects with common remaining consumption processes are grouped into remote sensing data governance units. Then, by combining the capacity status of various heterogeneous storage sources and the existing data distribution, a target aggregation storage source is determined, allowing related data within the same governance unit to form temporary collaborative copies while retaining their original effective physical storage locations. After the production process is completed, the remote sensing data governance unit is further reconstructed based on the updated set of remaining consumption processes and newly generated intermediate data. A new target aggregation storage source is then determined for the reconstructed remote sensing data governance unit, allowing newly added or recombined logical data objects to re-form corresponding common storage locations. Finally, the original temporary collaborative copies are converged based on the newly determined target aggregation storage source, ensuring that data organization, storage location, and the actual calling relationship of production processes maintain a dynamic correspondence. Attached Figure Description

[0008] Figure 1 This is a flowchart of a remote sensing big data governance method that supports heterogeneous storage according to the present invention; Figure 2 This is a schematic diagram illustrating the construction and relationship breakdown of the remote sensing data governance unit of the present invention; Figure 3 This is a schematic diagram of the collaborative aggregation of remote sensing data governance units according to the present invention; Figure 4 This is a schematic diagram of the remote sensing data governance unit reconstruction, re-aggregation, and temporary collaborative replica convergence of the present invention; Figure 5 This is a schematic diagram of the structure of a remote sensing big data governance system that supports heterogeneous storage according to the present invention. Detailed Implementation

[0009] The present invention will be further described below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described are only used to illustrate the technical solutions of the present invention, and are not intended to limit the scope of protection of the present invention. Those skilled in the art can adjust some technical features, processing order or module connection relationship according to actual application needs without departing from the concept of the present invention, and can also combine related technical features in different embodiments. All technical solutions formed should fall within the protection scope of the present invention.

[0010] It should be noted that, unless otherwise expressly defined, the technical terms used in this specification should be interpreted according to their common understanding by those skilled in the art. The terms "first," "second," etc., used in this specification are only used to distinguish corresponding technical objects and do not indicate a fixed order, quantitative relationship, or degree of importance between the stated technical objects. Terms such as "include" and "comprising" indicate that, in addition to the technical features, steps, or modules listed subsequently, other technical features, steps, or modules may also be included. The term "connection" in this specification includes direct connections and indirect connections, and may represent spatial connection relationships, data transmission relationships, or function call relationships.

[0011] like Figure 1 As shown, a remote sensing big data governance method supporting heterogeneous storage includes the following steps: Acquire heterogeneous storage sources, remote sensing production data, production processes and process input-output relationships, establish logical data objects that include spatial processing range, data capacity and effective physical storage location set, and establish consumption relationships based on process input-output relationships; Obtain the execution status of the production process, determine the set of remaining consumption processes based on the consumption relationship and the execution status, and determine the dual coupling quantity based on the spatial processing range of any two logical data objects and the set of remaining consumption processes, and establish a data governance coupling edge; Candidate connected components are formed based on the data governance coupling edges, and the candidate connected components are split according to the common remaining consumption process to obtain a set of remote sensing data governance units. Based on the available storage capacity of heterogeneous storage sources, the data capacity of logical data objects, and the set of effective physical storage locations, the target aggregation storage source for each remote sensing data governance unit is determined. Logical data objects not located in the target aggregation storage source are copied to the target aggregation storage source to form temporary collaborative copies, and the set of valid physical storage locations is updated. After the production process is completed, the set of remaining consumption processes is updated, new logical data objects are created for the intermediate data generated by the production process, and the data governance coupling edge and remote sensing data governance unit are redefined. Based on the redefined remote sensing data governance unit, the target aggregation storage source is redefined, and logical data objects not located in the target aggregation storage source are copied to the target aggregation storage source to form a new temporary collaborative copy. The original temporary collaborative copy is then converged, and the set of effective physical storage locations is updated.

[0012] In one embodiment of the present invention, the following steps are first taken to obtain the heterogeneous storage sources, remote sensing production data, production processes, and process input-output relationships currently connected to the remote sensing production system. The heterogeneous storage sources include network-attached storage, object storage, and distributed storage that differ in storage type or access protocol but are connected to the same remote sensing production system. The storage source identifier, storage type, access protocol, access interface, physical storage location, total storage capacity, available storage capacity, and current writable status of each heterogeneous storage source are read, and corresponding storage resource records are established according to a unified data structure. The storage resource records retain the original access methods of each heterogeneous storage source, while concentrating the information needed to subsequently determine the data location and storage status into a single record structure, enabling data locations of different storage types to be accessed according to a unified data object.

[0013] Logical data objects are established for the raw remote sensing images, control data, and intermediate data generated during the production process that enter the remote sensing production system. During establishment, a corresponding logical data identifier is assigned to each piece of remote sensing production data, and the spatial processing range, data type, data capacity, generation process, data version, and effective physical storage location set of that piece of remote sensing production data are associated with the logical data identifier. The spatial processing range is obtained based on the work grid, production slab, or other predetermined spatial units already determined for the current remote sensing production task, and is used to represent the spatial area corresponding to that piece of remote sensing production data in the current production task; the data capacity is the amount of data occupied by the corresponding remote sensing production data in the storage medium; the data version is used to distinguish different versions of the same logical data object generated during the production process.

[0014] The set of valid physical storage locations consists of physical storage locations that can currently access the corresponding logical data objects. Each physical storage location is associated with a heterogeneous storage source. When a piece of remote sensing production data is stored in only one heterogeneous storage source, the corresponding physical storage location is written into the set of valid physical storage locations. When the same piece of remote sensing production data exists in multiple heterogeneous storage sources, all physical storage locations are written into the set of valid physical storage locations. The logical data identifier remains unchanged, and the addition or removal of physical storage locations is recorded through the set of valid physical storage locations, ensuring that the logical data object and its actual storage location maintain an independent correspondence. In this way, when performing cross-storage replication of remote sensing production data, only the corresponding set of valid physical storage locations needs to be updated, without re-establishing the logical data object.

[0015] After obtaining the logical data objects, the configured production processes and their input-output relationships within the current remote sensing production task are read. Each production process is organized according to the processing order within the remote sensing production task. The input-output relationships record the input data called by each production process and the output data generated after execution. When the output data of one production process continues to serve as the input data for another, the two production processes are connected according to the corresponding data transfer relationship, forming a remote sensing production process dependency topology. This allows the determination of the data transfer order between different production processes and further mapping it to each logical data object.

[0016] The consumption relationship between logical data objects and production processes is determined one by one based on the input-output relationship of each process. For any logical data object, the input configuration of its corresponding remote sensing production data in the current remote sensing production task is read. When the remote sensing production data is configured as input data for a certain production process, a consumption relationship is determined between the logical data object and that production process. When the same logical data object is configured as input data by multiple production processes, a consumption relationship is established between the logical data object and each production process respectively. No consumption relationship is established between the logical data object and a production process that does not configure it as input data. The consumption relationship is directly derived from the process input-output relationship already determined in the current remote sensing production task, corresponding to the actual data call relationship of the production process to the logical data object.

[0017] After the above data organization, heterogeneous storage sources form corresponding storage resource records, remote sensing production data forms corresponding logical data objects, and the process input-output relationship further forms the consumption relationship between the logical data object and the production process. The spatial processing range, data capacity, and effective physical storage location set in the logical data object are retained as the basic data for subsequent governance processes, while the consumption relationship, together with the execution status of the production process, participates in the determination of the remaining consumption process set.

[0018] Furthermore, the execution status of each production process is obtained, and the previously established consumption relationships and the spatial processing range corresponding to each logical data object are invoked. For any logical data object, the production processes that have a consumption relationship with that logical data object are read along its consumption relationship. Production processes that are not yet completed are retained, while those that have been completed are excluded. The resulting production processes together constitute the set of remaining consumption processes for that logical data object. Since the consumption relationship reflects the pre-determined data calling relationship of the current remote sensing production task, and the execution status reflects the actual progress of the production process, the set of remaining consumption processes changes with the actual execution of the production processes. This set is used to retain information on which production processes will still call the logical data object, while production processes that have completed data consumption no longer participate in the determination of subsequent coupling relationships.

[0019] After obtaining the set of remaining consumption processes for each logical data object, determine the spatial correlation and consumption correlation for any two logical data objects. Let logical data object a and logical data object b be logical data object a and logical data object b, respectively, and their spatial processing ranges be respectively... and First, determine the overlapping area of ​​the two spatial processing ranges and calculate the area of ​​the overlapping area; then compare the areas of the two spatial processing ranges and select the smaller area as the calculation benchmark for spatial association.

[0020] The spatial association between logical data object a and logical data object b is represented as follows: ,in, This represents the spatial association quantity between logical data object a and logical data object b. This represents the area of ​​the overlapping region between two spatial processing ranges. and These represent the spatial processing area of ​​the two logical data objects, respectively. This represents a preset area correction value, using the same unit of measurement as the spatial processing range area. It is used to maintain the validity of the calculation relationship when the spatial processing range area is zero or close to zero. By using a smaller spatial processing range as a comparison benchmark, the spatial correlation quantity reflects the degree of coverage of two logical data objects within a smaller spatial range, rather than simply representing the spatial relationship by the absolute size of the overlapping area.

[0021] The consumption correlation quantity is determined based on the remaining consumption process sets of two logical data objects, let logical data object a and logical data object b be respectively... and First, determine the intersection of the two sets of remaining consumption processes to obtain the production processes that are common to both. Then, determine the union of the two sets of remaining consumption processes to obtain all the production processes that both still need to participate in.

[0022] The consumption association between two logical data objects is represented as follows: ,in, This represents the consumption association quantity between logical data object a and logical data object b. This represents the number of production processes commonly contained in two sets of remaining consumption processes. This represents the number of production processes included after merging two sets of remaining consumption processes. This represents a preset quantity correction value, which is a positive correction value corresponding to the number of production processes. The more remaining production processes that two logical data objects need to participate in together, and the fewer other production processes they participate in independently, the greater the resulting consumption correlation quantity will be, so that the calculation result can correspond to the degree of joint consumption between the two under the current production progress.

[0023] Given that the spatial correlation and consumption correlation are already determined, the spatial correlation and consumption correlation of two logical data objects in the same group are multiplied together, and the resulting product is determined as the dual coupling quantity between the two logical data objects. The dual coupling quantity is influenced by both the overlapping relationship of spatial processing ranges and the common relationship of remaining consumption processes. When two logical data objects only have spatial overlap but no common remaining consumption processes, the corresponding consumption correlation cannot form effective common consumption support; when two logical data objects have the same or partially the same remaining consumption processes, but their spatial correlation is low, the spatial correlation restricts the dual coupling quantity. Therefore, the dual coupling quantity is not determined by independently judging the two correlation dimensions, but rather by combining the spatial processing relationship with the remaining consumption relationship under the current production state into a single governance correlation quantity, used to subsequently determine whether a data governance coupling edge is established between the logical data objects.

[0024] After determining the dual coupling amount, the remaining consumption process sets corresponding to the two logical data objects are read, and the intersection of the two remaining consumption process sets is determined. When the intersection is not empty, that is, when the two logical data objects have at least one common remaining consumption process, the dual coupling amount between the two logical data objects is compared with a preset governance coupling threshold. When the dual coupling amount is not lower than the preset governance coupling threshold, a data governance coupling edge is established between the two logical data objects. When the intersection of the two remaining consumption process sets is empty, or the dual coupling amount is lower than the preset governance coupling threshold, no data governance coupling edge is established between the two logical data objects. The common remaining consumption process is used to limit the relationship that the two logical data objects will continue to be called by the same production process, while the dual coupling amount is used to characterize the comprehensive correlation between the two in terms of spatial processing relationship and remaining consumption relationship. The preset governance coupling threshold is used as the comparison benchmark for establishing the data governance coupling edge, and is preset according to the data organization granularity of the remote sensing production task, without changing the determination process of the dual coupling amount.

[0025] After traversing the logical data object combinations requiring association judgment according to the above processing, each logical data object maintains its corresponding set of remaining consumption processes, forming a dual coupling quantity between any two logical data objects and a data governance coupling edge that meets the establishment conditions. The data governance coupling edge records the connection relationship between logical data objects that still have a common consumption basis and meet the dual coupling condition under the current production process execution state. As completed production processes are removed from the set of remaining consumption processes, the consumption association quantity and dual coupling quantity between the same logical data objects can be re-determined according to the actual production state, so that the formed governance relationship corresponds to the currently incomplete data consumption process. The obtained data governance coupling edge and the set of remaining consumption processes continue to serve as the data basis for forming candidate connected components and verifying common remaining consumption processes.

[0026] Furthermore, based on the aforementioned determined data governance coupling edges, the set of remaining consumption processes corresponding to each logical data object, and the dual coupling amount between logical data objects, a remote sensing data governance relationship graph is constructed to represent the association relationships of logical data objects in the current production state. All logical data objects are treated as nodes in the remote sensing data governance relationship graph. When a data governance coupling edge has been established between two logical data objects, the data governance coupling edge is retained as a connection between the corresponding two nodes. Logical data objects without established data governance coupling edges do not form direct connections. The connection relationships recorded in the remote sensing data governance relationship graph originate from the data governance coupling edges already determined in the current production process execution state. Therefore, when the execution state of the production process changes and the data governance coupling edges are redefined, the connection relationships in the remote sensing data governance relationship graph are also redefined accordingly.

[0027] After forming the remote sensing data governance relationship graph, candidate connected components are identified based on the connectivity between nodes. For any candidate connected component, the logical data objects it contains can be interconnected along one or more data governance coupling edges, while there are no data governance coupling edges between the logical data objects in this candidate connected component and the logical data objects in other candidate connected components. The candidate connected component only reflects the graph connectivity formed by logical data objects under the action of data governance coupling edges, and is not yet directly used as a remote sensing data governance unit. It is necessary to further determine whether it has a process basis that can jointly participate in subsequent production based on the remaining consumption process set of each logical data object within it.

[0028] For each candidate connected component, the set of remaining consumption steps corresponding to all logical data objects within it is read, and the intersection of each set of remaining consumption steps is determined sequentially. The production steps retained in the intersection are the common remaining consumption steps of all logical data objects in the candidate connected component. When the intersection is not empty, it indicates that each logical data object in the candidate connected component will be called by at least one incomplete production step. In this case, the existing data governance coupling edges within the candidate connected component are maintained, and the candidate connected component is determined as a remote sensing data governance unit. The common remaining consumption steps corresponding to this remote sensing data governance unit are recorded. In this way, the remote sensing data governance unit not only has the connection relationship formed by the data governance coupling edges, but also has the common production step call relationship that runs through all logical data objects within the unit.

[0029] While nodes within a candidate connected component are interconnected, this does not necessarily mean that all logical data objects share a common remaining consumption process. For example, some logical data objects in a candidate connected component may establish data governance coupling edges through a production process, while other logical data objects may form connections with intermediate logical data objects through another production process. In this case, the entire candidate connected component remains connected in the remote sensing data governance relationship graph, but the sets of remaining consumption processes for all logical data objects may not have a common intersection. For such candidate connected components, instead of directly identifying them as remote sensing data governance units, they undergo an association relationship decomposition process, using common remaining consumption processes as a verification condition for whether to continue overall governance. This process constrains both the indirect connections on the graph and the common invocation relationships of actual production processes, preventing logical data objects without common subsequent consumption relationships from being fixed within the same governance scope simply because multiple local connections appear consecutively.

[0030] like Figure 2 As shown, multiple logical data objects in a candidate connected component are connected through data governance coupling edges. However, this connectivity does not directly indicate that all logical data objects share a common remaining consumption process. When a candidate connected component does not share a common remaining consumption process, the double coupling amount corresponding to each data governance coupling edge within the candidate connected component is read, and the data governance coupling edge with the smallest double coupling amount is deleted. After the data governance coupling edge is deleted, the candidate connected components are re-identified based on the remaining connectivity, and the common remaining consumption process for each candidate connected component is determined. Figure 2 In the process, the original candidate connected components are divided into two separate connected parts after the data governance coupling edge with the smallest double coupling is deleted. Each connected part has a common remaining consumption process, and therefore they are determined as remote sensing data governance units.

[0031] After deletion, candidate connected components are re-identified based on the remaining data governance coupling edges. It is not assumed that a single deletion will necessarily create a fixed number of new candidate connected components; instead, the actual node connection relationships formed after deletion are used as the standard. For each re-identified candidate connected component, the remaining consumption process set of all logical data objects within it is read again, and the common remaining consumption process is re-determined.

[0032] If the redefined candidate connected components already share a common remaining consumption process, then the further splitting of the candidate connected component is stopped, and it is identified as a remote sensing data governance unit. If there is still no common remaining consumption process and there are still data governance coupling edges within the candidate connected component, then the double coupling amount corresponding to each data governance coupling edge is compared, the data governance coupling edge with the smallest double coupling amount is deleted, and the connectivity relationship between nodes is re-identified. If there is no common remaining consumption process and there are no more data governance coupling edges within the candidate connected component, then the further splitting of the candidate connected component is stopped, and it is not identified as a remote sensing data governance unit.

[0033] After obtaining the set of remote sensing data governance units, the correspondence between the logical data objects contained in each unit, the corresponding data governance coupling edges, and the common remaining consumption processes is preserved. The original data capacity and effective physical storage location set of each logical data object do not change due to the formation of the governance unit, but continue to be associated with the logical data object. In this way, when determining the target aggregation storage source according to the remote sensing data governance unit, the remote sensing data governance unit can be directly used as the overall processing object. The data capacity and effective physical storage location set of each logical data object within the unit can be read, and the corresponding storage aggregation location can be determined by combining the storage status of heterogeneous storage sources.

[0034] Furthermore, based on the established set of remote sensing data governance units, the data capacity and effective physical storage location set of the logical data objects contained in each remote sensing data governance unit are read. Simultaneously, the total storage capacity, available storage capacity, current writable status, and access interface are read from the storage resource records corresponding to each heterogeneous storage source. Since the remote sensing data governance units have already been formed according to a common residual consumption process in the aforementioned processing, this stage no longer selects storage locations independently for each logical data object. Instead, a single remote sensing data governance unit is used as the processing object for storage aggregation, comparing the capacity and current data distribution of different heterogeneous storage sources for that remote sensing data governance unit.

[0035] For any remote sensing data governance unit, candidate storage sources that can participate in aggregation are first screened from all heterogeneous storage sources. The current writable status of each heterogeneous storage source is read; heterogeneous storage sources that are currently not writable are not included in subsequent comparisons. For heterogeneous storage sources that remain writable, their access interfaces in the storage resource records are used to determine whether they support access to the corresponding remote sensing production data; heterogeneous storage sources that cannot provide the corresponding access interface are excluded. After completing the above screening, based on the effective physical storage location set of each logical data object within the remote sensing data governance unit, it is determined which logical data objects are already located in the current heterogeneous storage source and which are not, and the data capacity of the logical data objects not yet located in the current heterogeneous storage source is calculated. If the available storage capacity of the current heterogeneous storage source cannot accommodate the data that needs to be newly copied for these logical data objects, then that heterogeneous storage source is not retained as a candidate storage source. Through continuous screening based on writable status, access interface, and available storage capacity, candidate storage sources for the corresponding remote sensing data governance unit are obtained.

[0036] When the selected candidate storage sources are empty, no target aggregation storage source is determined for the current remote sensing data governance unit, and the corresponding temporary collaborative copy formation process is not executed. Each logical data object maintains its current set of effective physical storage locations and logical location mapping relationship, and the production process continues to access data based on the existing effective physical storage locations. When the current writable status, access interface support, or available storage capacity of heterogeneous storage sources is updated, candidate storage source selection is re-executed for the remote sensing data governance unit. If the candidate storage sources are not empty, the capacity of already stored data, the capacity of newly copied data, and the aggregation evaluation value are then determined.

[0037] For each candidate storage source, the existing data capacity and the newly copied data capacity are determined according to the current actual distribution status of the logical data objects. Specifically, the logical data objects within the remote sensing data governance unit are traversed. If the set of valid physical storage locations for a logical data object contains a physical storage location corresponding to the current candidate storage source, the data capacity of that logical data object is included in the existing data capacity. If the set of valid physical storage locations does not contain a physical storage location corresponding to the current candidate storage source, the data capacity of that logical data object is included in the newly copied data capacity. The data capacities of the former type of logical data objects are accumulated to obtain the existing data capacity corresponding to the current candidate storage source, and the data capacities of the latter type of logical data objects are accumulated to obtain the newly copied data capacity corresponding to the current candidate storage source. The existing data capacity reflects the scale of data already in the remote sensing data governance unit and located in the current candidate storage source. The newly copied data capacity corresponds to the increased storage space required when aggregating the remote sensing data governance unit to the current candidate storage source. Both are directly derived from the existing data capacity and the set of valid physical storage locations of the logical data objects and do not change the data attributes of the logical data objects themselves.

[0038] The remaining capacity ratio of candidate storage sources after the remote sensing data governance unit completes aggregation is further determined based on the newly copied data capacity. Specifically, for any candidate storage source, the newly copied data capacity corresponding to the current remote sensing data governance unit is subtracted from its current available storage capacity to obtain the remaining storage capacity after aggregation. This remaining storage capacity is then compared with the total storage capacity of the candidate storage source, and the ratio is determined as the remaining capacity ratio after aggregation. This remaining capacity ratio reflects the remaining capacity status of the candidate storage source after the current remote sensing data governance unit completes the new copying. When the remaining storage capacity after aggregation is lower than a preset storage source reserved capacity threshold, the corresponding candidate storage source is excluded from the candidate range of the current remote sensing data governance unit, ensuring that subsequent aggregation evaluation is only performed among candidate storage sources that meet the reserved capacity requirements.

[0039] For each candidate storage source that meets the capacity requirements, the aggregation evaluation value is determined based on the proportion of existing data capacity, newly replicated data capacity, and remaining capacity. The aggregation evaluation value can be expressed as: ,in, Indicates the first Each remote sensing data governance unit corresponds to the aggregation evaluation value of the m-th candidate storage source. This represents the capacity of data already stored in the m-th candidate storage source within the remote sensing data governance unit. This represents the additional replicated data capacity required for the remote sensing data governance unit to aggregate to the m-th candidate storage source. This represents the remaining capacity percentage of the m-th candidate storage source after aggregation. This represents the preset capacity correction value, which uses the same storage capacity unit as the stored data capacity and the newly copied data capacity. It is used to maintain the continuity of the calculation relationship of the aggregation evaluation value when the corresponding data capacity is zero. The larger the stored data capacity is relative to the newly copied data capacity, the more data belonging to the remote sensing data governance unit is already in the current candidate storage source, and the less data needs to be copied. The remaining capacity ratio allows this data distribution relationship to be compared together with the capacity status of the candidate storage source after aggregation.

[0040] The same aggregation evaluation value for each candidate storage source corresponding to a remote sensing data governance unit is calculated, and the obtained aggregation evaluation values ​​are compared. The candidate storage source with the largest aggregation evaluation value is determined as the target aggregation storage source for that remote sensing data governance unit. The above determination process is performed on each remote sensing data governance unit in the set of remote sensing data governance units to form a correspondence between each remote sensing data governance unit and the target aggregation storage source.

[0041] Each remote sensing data governance unit determines its target aggregation storage source. Once the target aggregation storage source for the current remote sensing data governance unit is determined, the corresponding new replication data capacity for that unit is recorded. When determining the target aggregation storage source for the next remote sensing data governance unit, for heterogeneous storage sources already identified as target aggregation storage sources by previous units, the capacity of new replication data already determined to be written to that heterogeneous storage source is deducted from its current available storage capacity. This deducted capacity is then used for candidate storage source selection in the next remote sensing data governance unit. This process is repeated to determine the target aggregation storage source for each remote sensing data governance unit in the set, ensuring that the capacity used by subsequent governance units includes the new replication data already determined by previous governance units.

[0042] Once the target aggregation storage source is determined, the original set of valid physical storage locations for each logical data object remains unchanged. This determination result serves as the target storage location for data replication of logical data objects that are not yet located in the target aggregation storage source within the remote sensing data governance unit.

[0043] Furthermore, when the common remaining consumption process corresponding to the remote sensing data governance unit enters the input preparation stage, the target aggregation storage source corresponding to the remote sensing data governance unit is read, and the logical data identifiers, effective physical storage location sets, and access protocols and interfaces of relevant heterogeneous storage sources for each logical data object are invoked. The target aggregation storage source has already been determined based on the data distribution and storage capacity status within the remote sensing data governance unit. At this stage, the storage location is not re-evaluated; instead, logical data objects within the remote sensing data governance unit that are not yet located in the target aggregation storage source are collaboratively placed according to the determination result. For logical data objects whose effective physical storage location set already includes a physical storage location corresponding to the target aggregation storage source, their existing storage status is maintained. For logical data objects whose effective physical storage location set does not contain a physical storage location corresponding to the target aggregation storage source, the readable data location is determined based on the current effective physical storage location, and their remote sensing production data is read according to the access protocols and interfaces of the corresponding heterogeneous storage sources before being written to the target aggregation storage source. In this way, whether data within the same remote sensing data governance unit needs to be copied is determined by its current physical distribution, and duplicate copies are not generated for data that is already located in the target aggregation storage source.

[0044] like Figure 3 As shown, logical data objects in a remote sensing data governance unit can be located in different heterogeneous storage sources. Figure 3 In this scenario, logical data objects A, C, and E belong to the same remote sensing data governance unit. Logical data objects A and C are located in heterogeneous storage source 1 and 2, respectively, while logical data object E is already located in heterogeneous storage source 3. When heterogeneous storage source 3 is designated as the target aggregation storage source for this remote sensing data governance unit, the existing physical storage location of logical data object E remains unchanged. Data is then read from the current valid physical storage locations of logical data objects A and C and copied to heterogeneous storage source 3, forming temporary collaborative copies A and C, respectively. During the copying process, the original valid physical storage locations of logical data objects A and C are retained, ensuring that the target aggregation storage source adds corresponding physical storage locations without altering the already valid storage locations before copying.

[0045] During data replication, the original logical data identifiers of logical data objects remain unchanged, while the physical storage locations already recorded in the set of valid physical storage locations before replication are retained, ensuring that the data in the original storage locations remains in a valid access state. The new copy formed by writing to the target aggregation storage source is marked as a temporary collaborative copy. Temporary collaborative copies are added during the collaborative placement of remote sensing data governance units to meet the data aggregation needs of the current common remaining consumption processes. They do not immediately replace the original data locations after formation, nor are they directly added to the set of valid physical storage locations before the writing process is complete. Production processes still obtain the corresponding data according to the original logical data identifiers. The correspondence between logical data identifiers and the original valid physical storage locations remains valid during replication, thereby separating data writing across different access protocols from data retrieval by upper-level production processes.

[0046] After the temporary collaborative copy is written, an integrity consistency check is performed on the source copy and the temporary collaborative copy that provided the data before replication. The integrity consistency check verifies the correspondence between the data content represented by the source copy and the temporary collaborative copy. Only when the check results match is the temporary collaborative copy confirmed as a valid data location for the logical data object. If the integrity consistency check is not yet complete, the original set of valid physical storage locations remains unchanged, and the production process continues to access data based on the original location relationships, preventing the temporary collaborative copy, which is still being written, from prematurely entering the valid access range. After the integrity consistency check passes, the physical storage location corresponding to the target aggregation storage source is added to the set of valid physical storage locations for the logical data object. At this time, the original physical storage location is still retained, enabling a logical data object to be associated with both the valid physical storage location before replication and the temporary collaborative copy location in the target aggregation storage source.

[0047] After the set of valid physical storage locations is updated, a new logical location mapping relationship is established based on the logical data identifier and the updated set of valid physical storage locations. The logical location mapping relationship records the correspondence between the logical data identifier and each physical storage location that can currently access the logical data object. This allows the production process to determine the actual access location based on the currently effective logical location mapping relationship when initiating data calls through the logical data identifier. When establishing a new logical location mapping relationship, the original logical location mapping relationship is retained until the new logical location mapping relationship is established. After the new logical location mapping relationship is established, it is used as the currently effective logical location mapping relationship, allowing newly added target aggregation storage source locations to enter the subsequent data access scope. The logical data identifier does not change during this process; what changes is the valid physical storage location it can map to. Therefore, adding physical data locations does not require synchronously changing the data reference relationships already configured in the production process.

[0048] For multiple logical data objects within the same remote sensing data governance unit, the decision to perform replication is determined based on their respective sets of valid physical storage locations. After the corresponding temporary collaborative copy passes integrity and consistency verification, the sets of valid physical storage locations are updated. Once all logical data objects requiring replication have completed the above processing, the data in this remote sensing data governance unit, originally scattered across different heterogeneous storage sources, acquires valid physical storage locations pointing to the same target aggregation storage source. The original logical data identifiers and other valid physical storage locations of each logical data object are still recorded. The target aggregation storage source thus becomes a shared physical storage location that can be accessed by the remaining consumption processes when reading data from this remote sensing data governance unit, while the original storage locations continue to exist as existing valid locations for the logical data objects.

[0049] After a temporary collaborative copy is formed, its corresponding physical storage location is retained as a new location in the set of valid physical storage locations at the current stage. However, the relationship between the temporary collaborative copy and the original copy is not permanent. Its marked state continues to be associated with the corresponding logical data object and remote sensing data governance unit, so that when subsequent production processes are completed, the set of remaining consumption processes changes, and the remote sensing data governance unit is re-determined, it is possible to identify which valid physical storage locations are temporary collaborative copies formed by the collaborative placement of this governance unit. The state of the temporary collaborative copy formed after collaborative placement, the updated set of valid physical storage locations, and the currently effective logical location mapping relationship together constitute the basis for subsequent adjustments to the data storage state based on the actual execution results of the production processes.

[0050] Upon completion of any production process, its completion status is obtained, and the logical data objects consumed by that process are determined based on established consumption relationships. For each logical data object consumed by that process, the completed production processes are removed from its remaining consumption process set, ensuring that the set only contains processes still requiring the logical data object to be read. This update directly uses the actual completion result of the production process without predicting whether a process is about to complete. Therefore, the remaining consumption status of logical data objects can change synchronously with the actual progress of the remote sensing production task. For logical data objects not consumed by the current production process, their remaining consumption process set remains unchanged and is not altered by the completion of other production processes.

[0051] Furthermore, when intermediate data is generated after the production process is completed, a new logical data object is created for the intermediate data according to the aforementioned method for establishing logical data objects. This object records the spatial processing range, data type, data capacity, generation process, data version, and effective physical storage location set corresponding to the intermediate data, and assigns a corresponding logical data identifier. Based on the production process and its input-output relationship, the set of remaining consumption processes corresponding to the new logical data object is determined from the production processes that have not yet been completed after the current production process and require this intermediate data as input data. Therefore, the newly generated intermediate data is not stored as data independent of the existing governance process, but rather enters the data consumption relationship of the current remote sensing production task as a new logical data object. Its spatial processing range and the set of remaining consumption processes continue to participate in determining the data governance coupling relationship.

[0052] Based on the updated set of remaining consumption processes and the new logical data objects, the dual coupling quantities between related logical data objects are redefined. The determination process follows the aforementioned handling relationships for spatial and consumption correlation quantities. Spatial correlation quantities are determined based on the current spatial processing range of the logical data objects, and consumption correlation quantities are determined based on the updated set of remaining consumption processes. The dual coupling quantities are then determined using both. For logical data objects that satisfy common remaining consumption process conditions and governance coupling conditions, new data governance coupling edges are established; existing data governance coupling edges that no longer meet these conditions are not retained. Since production processes have been removed from the relevant set of remaining consumption processes, the original consumption correlation quantities between logical data objects may change accordingly. Newly generated intermediate data can also form new data governance coupling relationships with existing logical data objects based on its spatial processing range and the set of remaining consumption processes.

[0053] Candidate connected components are formed based on the newly determined data governance coupling edges, and remote sensing data governance units are redefined based on whether the logical data objects within each candidate connected component share common remaining consumption processes. If the logical data objects in the original remote sensing data governance unit still share common remaining consumption processes, the corresponding logical data objects can continue to form a reconstructed remote sensing data governance unit. If the common remaining consumption relationship in the original remote sensing data governance unit has disappeared, the data governance coupling edges with smaller double coupling amounts are decoupled according to the aforementioned candidate connected component splitting method, and the common remaining consumption processes of each candidate connected component are re-verified until a reconstructed remote sensing data governance unit that meets the conditions for establishing a governance unit is formed. For new logical data objects generated by the current production process, when they form a data governance coupling edge that meets the conditions with existing logical data objects, they participate in the reconstruction of the remote sensing data governance unit together with the corresponding logical data object; when they do not form a relationship that meets the conditions with existing remote sensing data governance units, whether to form a new remote sensing data governance unit is determined based on their actual connection relationship with other logical data objects. In this way, the composition of remote sensing data governance units is not fixed at the beginning of the production task, but is redefined as the actual remaining consumption relationship decreases and intermediate data is generated.

[0054] like Figure 4 As shown, after the production process is completed, the set of remaining consumption processes of the relevant logical data objects is updated according to the actual consumption status. At the same time, the intermediate data generated by the production process is established as a new logical data object and participates in the reconstruction of the remote sensing data governance unit. Figure 4 In the process, logical data objects A, B, and C, which were previously located within the same remote sensing data governance unit, undergo a re-determination of their governance relationships based on the updated set of remaining consumption processes after the production process is completed. Logical data objects B, C, and a new logical data object D then form the reconstructed remote sensing data governance unit. For the reconstructed remote sensing data governance unit, a new target aggregation storage source is determined, and new temporary collaborative copies are created for logical data objects not yet located within the newly determined target aggregation storage source. Temporary collaborative copies created before reconstruction whose physical storage locations correspond to the newly determined target aggregation storage source are retained; those that do not correspond but have already formed valid physical storage locations corresponding to the newly determined target aggregation storage source are deleted. For logical data object A, whose remaining consumption process set is empty, its temporary collaborative copy A is deleted, and the set of valid physical storage locations is updated accordingly.

[0055] After obtaining the reconstructed remote sensing data governance units, the data capacity and current effective physical storage location set of the logical data objects contained in each reconstructed remote sensing data governance unit are read, and the available storage capacity, current writable status, and access interface of each heterogeneous storage source are reread. Following the aforementioned process for determining the target aggregation storage source, candidate storage sources are re-screened for each reconstructed remote sensing data governance unit, the stored data capacity and newly copied data capacity are determined, and the corresponding target aggregation storage source is re-determined based on the current data distribution and storage capacity status. When the reconstructed remote sensing data governance unit has the same composition as the pre-reconstruction remote sensing data governance unit, its target aggregation storage source is still re-determined according to the current storage status; when the governance unit is split, recombined, or new logical data objects are added, the re-determined target aggregation storage source is used to form a new common data access location for the current governance unit.

[0056] When the common remaining consumption process corresponding to the reconstructed remote sensing data governance unit enters the input preparation stage, the newly determined target aggregation storage source is read. For logical data objects in the effective physical storage location set that already have a physical storage location corresponding to the target aggregation storage source, their current storage state is maintained; for logical data objects that do not have a corresponding physical storage location, they are copied to the newly determined target aggregation storage source according to the access protocol and access interface of the corresponding heterogeneous storage source, forming a new temporary collaborative copy. After the new temporary collaborative copy is written and passes the integrity and consistency check, the corresponding physical storage location is added to the effective physical storage location set, and the logical location mapping relationship is updated.

[0057] After the reconstructed collaborative placement is completed, the temporary collaborative copies formed before the reconstruction are converged. For logical data objects with an empty set of remaining consumption processes, their temporary collaborative copies and corresponding physical storage locations are deleted. For logical data objects with a non-empty set of remaining consumption processes that belong to the reconstructed remote sensing data governance unit, it is determined whether the physical storage location corresponding to the original temporary collaborative copy corresponds to the newly determined target aggregation storage source. If they correspond, they are retained; if they do not correspond but a valid physical storage location corresponding to the newly determined target aggregation storage source has already been formed, the original temporary collaborative copy is deleted. For logical data objects that do not belong to the reconstructed remote sensing data governance unit but have other valid physical storage locations, the original temporary collaborative copy is deleted. The valid physical storage location set and logical location mapping relationship are updated based on the copy retention or deletion results.

[0058] As subsequent production processes continue, the updated set of remaining consumption processes, the reconstructed remote sensing data governance units, and the updated set of effective physical storage locations are directly used as the new governance state. After each subsequent production process is completed, the set of remaining consumption processes for the corresponding logical data objects is reduced again based on the actual execution results. New logical data objects are then created for the newly generated intermediate data. The data governance coupling edges and remote sensing data governance units are redefined, and the target aggregation storage source corresponding to the reconstructed remote sensing data governance units is redefined and collaboratively positioned. Finally, the temporary collaborative copy before reconstruction is converged based on the redefined target aggregation storage source. Thus, the actual execution results of the production processes can continuously feed back to the remaining consumption state, data governance coupling relationship, and storage location relationship of the logical data objects, enabling the remote sensing data governance units to be formed, maintained, split, or recombined as the production process progresses. Simultaneously, the temporary collaborative copy formed for phased joint consumption is updated as the joint consumption relationship changes.

[0059] In another embodiment of the present invention, a remote sensing big data governance system supporting heterogeneous storage is also provided, such as... Figure 5 As shown, the system includes a basic relationship construction module, a coupling relationship determination module, a governance unit construction module, an aggregation storage determination module, a collaborative replica formation module, and a closed-loop reconstruction module. Each module transfers data according to the governance order of remote sensing production data. The basic relationship construction module acquires heterogeneous storage sources, remote sensing production data, production processes, and process input-output relationships to form logical data objects and consumption relationships. It then provides the spatial processing range, data capacity, and effective physical storage location set corresponding to the logical data objects to subsequent modules. The coupling relationship determination module acquires the execution status of production processes, forms a set of remaining consumption processes based on consumption relationships, and determines the dual coupling quantity and data governance coupling edge by combining the spatial processing range of the logical data objects.

[0060] The governance unit construction module receives the data governance coupling edges and the set of remaining consumption processes, forms candidate connected components, and splits these candidate connected components according to common remaining consumption processes, outputting a set of remote sensing data governance units. The aggregation storage determination module reads the set of remote sensing data governance units, the data capacity of logical data objects, the set of effective physical storage locations, and the available storage capacity of heterogeneous storage sources to determine the target aggregation storage source corresponding to each remote sensing data governance unit. The collaborative replica formation module performs replication on logical data objects not located in the target aggregation storage source, forming temporary collaborative replicas, and writes the newly added physical storage locations into the set of effective physical storage locations.

[0061] The closed-loop reconstruction module receives the completion status of the production process, the current set of remaining consumption processes, the intermediate data generated by the production process, temporary collaborative copies, and the set of effective physical storage locations. After the production process is completed, it updates the set of remaining consumption processes, creates new logical data objects for the intermediate data, and redetermines the data governance coupling edge and the remote sensing data governance unit. The reconstructed remote sensing data governance unit is then provided to the aggregation storage determination module. This module redetermines the target aggregation storage source based on the available storage capacity of the current heterogeneous storage source, the data capacity of the logical data object, and the set of effective physical storage locations. The collaborative copy formation module forms a new temporary collaborative copy based on the redetermined target aggregation storage source and updates the set of effective physical storage locations. The closed-loop reconstruction module then forms a temporary collaborative copy before the result convergence reconstruction based on the redetermined target aggregation storage source and the new temporary collaborative copy, allowing the updated governance state to continue as the system input during subsequent production process execution. The data objects, sets, and storage states already formed in the aforementioned method implementation are transmitted between the various modules of the system. The specific determination process is consistent with the corresponding method implementation and will not be repeated here.

[0062] The embodiments of the present invention have been described above, but the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make other forms of transformation or adjustment without departing from the concept of the present invention, and all such transformations or adjustments should fall within the protection scope of the present invention.

Claims

1. A remote sensing big data governance method supporting heterogeneous storage, characterized in that, Includes the following steps: Acquire heterogeneous storage sources, remote sensing production data, production processes and process input-output relationships, establish logical data objects that include spatial processing range, data capacity and effective physical storage location set, and establish consumption relationships based on process input-output relationships; Obtain the execution status of the production process, determine the set of remaining consumption processes based on the consumption relationship and the execution status, and determine the dual coupling quantity based on the spatial processing range of any two logical data objects and the set of remaining consumption processes, and establish a data governance coupling edge; Candidate connected components are formed based on the data governance coupling edges, and the candidate connected components are split according to the common remaining consumption process to obtain a set of remote sensing data governance units. Based on the available storage capacity of heterogeneous storage sources, the data capacity of logical data objects, and the set of effective physical storage locations, the target aggregation storage source for each remote sensing data governance unit in the remote sensing data governance unit set is determined. Logical data objects not located in the target aggregation storage source are copied to the target aggregation storage source to form temporary collaborative copies, and the set of valid physical storage locations is updated. After the production process is completed, the set of remaining consumption processes is updated, new logical data objects are created for the intermediate data generated by the production process, and the data governance coupling edge and remote sensing data governance unit are redefined. Based on the redefined remote sensing data governance unit, the target aggregation storage source is redefined, and logical data objects not located in the target aggregation storage source are copied to the target aggregation storage source to form a new temporary collaborative copy. The original temporary collaborative copy is then converged, and the set of effective physical storage locations is updated.

2. The remote sensing big data governance method supporting heterogeneous storage according to claim 1, characterized in that, The process of establishing the logical data object and the consumption relationship includes: A storage resource record is established for each heterogeneous storage source. The storage resource record includes storage source identifier, storage type, access protocol, access interface, physical storage location, total storage capacity, available storage capacity, and current writable status. For each piece of remote sensing production data, a logical data identifier is assigned, and the corresponding spatial processing range, data type, data capacity, generation process, data version, and set of effective physical storage locations are recorded to obtain a logical data object; A remote sensing production process dependency topology is established based on the production process and the process input-output relationship. When a logical data object is configured as input data for a production process, a consumption relationship is established between the logical data object and the production process.

3. The remote sensing big data governance method supporting heterogeneous storage according to claim 1, characterized in that, Determine the set of remaining consumption processes and the dual coupling quantities, including: From the production processes that have a consumption relationship with each logical data object, extract the production processes whose execution status is incomplete, and form the set of remaining consumption processes for the corresponding logical data object. For any two logical data objects, the ratio of the overlapping area of ​​their spatial processing ranges to the sum of the smaller of the two spatial processing range areas and a preset area correction value is determined as the spatial correlation quantity; the ratio of the number of production processes contained in the intersection of their remaining consumption process sets to the sum of the number of production processes contained in the union of their remaining consumption process sets and a preset quantity correction value is determined as the consumption correlation quantity. The product of the spatial association quantity and the consumption association quantity is determined as the dual coupling quantity between the two logical data objects.

4. The remote sensing big data governance method supporting heterogeneous storage according to claim 1, characterized in that, The process of establishing data governance coupling edges includes: Determine the intersection of the sets of remaining consumption processes corresponding to two logical data objects, and identify the production processes contained in the intersection as the common remaining consumption processes of the two logical data objects; When two logical data objects share a common remaining consumption process, and the amount of dual coupling between the two logical data objects is not lower than a preset governance coupling threshold, a data governance coupling edge is established between the two logical data objects; when there is no common remaining consumption process or the amount of dual coupling is lower than the preset governance coupling threshold, the data governance coupling edge is not established.

5. A remote sensing big data governance method supporting heterogeneous storage according to claim 1, characterized in that, The process of determining the set of remote sensing data governance units includes: A remote sensing data governance relationship graph is established using logical data objects as nodes and data governance coupling edges as the connection relationships between nodes, and candidate connected components are determined from the remote sensing data governance relationship graph. The intersection of the remaining consumption process sets of each logical data object within a candidate connected component is determined as the common remaining consumption process; when a common remaining consumption process exists, the corresponding candidate connected component is determined as a remote sensing data governance unit. When there are no common remaining consumption processes, if there are data governance coupling edges in the corresponding candidate connected components, delete the data governance coupling edge with the smallest double coupling, redetermine the candidate connected components and common remaining consumption processes, and repeat the above process for the redetermined candidate connected components; if there are no data governance coupling edges in the corresponding candidate connected components, do not determine the candidate connected components as remote sensing data governance units, and obtain a set of remote sensing data governance units.

6. The remote sensing big data governance method supporting heterogeneous storage according to claim 1, characterized in that, The target aggregation and storage sources for each remote sensing data governance unit are determined, including: For each remote sensing data governance unit, the target aggregation storage source is determined one by one, and heterogeneous storage sources that are currently unwritable or do not support the corresponding access interface are screened out from heterogeneous storage sources. For each heterogeneous storage source after screening, the data capacity of logical data objects in the effective physical storage location set of the remote sensing data governance unit that contain the physical storage location corresponding to the heterogeneous storage source is accumulated to obtain the stored data capacity. The data capacity of logical data objects in the effective physical storage location set that do not contain the physical storage location corresponding to the heterogeneous storage source is accumulated to obtain the newly copied data capacity. Heterogeneous storage sources whose available storage capacity cannot accommodate the newly copied data capacity are screened out to obtain candidate storage sources. The remaining capacity ratio is determined by the remaining capacity after deducting the newly copied data capacity from the available storage capacity of the candidate storage sources and the total storage capacity. The sum of the stored data capacity and the preset capacity correction value is used as the numerator, and the sum of the newly copied data capacity and the preset capacity correction value is used as the denominator. The product of the resulting ratio and the remaining capacity ratio is determined as the aggregation evaluation value. The candidate storage source with the largest aggregation evaluation value is determined as the target aggregation storage source. When determining the target aggregation storage source for the next remote sensing data governance unit, the capacity of newly replicated data that has been determined to be aggregated to each heterogeneous storage source is deducted from the current available storage capacity of the corresponding heterogeneous storage source, and the candidate storage source is selected based on the capacity after deduction.

7. A remote sensing big data governance method supporting heterogeneous storage according to claim 1, characterized in that, The process of forming temporary collaborative copies and updating the set of effective physical storage locations includes: When the common remaining consumption process corresponding to the remote sensing data governance unit enters the input preparation stage, the target aggregation storage source of the remote sensing data governance unit is obtained. For logical data objects that do not have a physical storage location corresponding to the target aggregation storage source in the set of effective physical storage locations, they are copied to the target aggregation storage source through the access protocol and access interface of the corresponding heterogeneous storage source to form a temporary collaborative copy. During the replication process, the logical data identifier and the original valid physical storage location remain unchanged. After the temporary collaborative copy is written, the integrity and consistency verification between the source copy and the temporary collaborative copy is performed. After verification, the physical storage location corresponding to the target aggregation storage source is added to the set of valid physical storage locations of the corresponding logical data object, and a new logical location mapping relationship is established based on the logical data identifier and the updated set of valid physical storage locations.

8. A remote sensing big data governance method supporting heterogeneous storage according to claim 1, characterized in that, The reconstruction process of the remote sensing data governance unit includes: Once a production process is completed, the production process is removed from the set of remaining consumption processes corresponding to the logical data object consumed by the production process, resulting in an updated set of remaining consumption processes. A new logical data object is created for the intermediate data generated by the production process, and the set of remaining consumption processes corresponding to the new logical data object is determined according to the process input-output relationship; Based on the updated set of remaining consumption processes and the set of remaining consumption processes corresponding to the new logical data objects, the dual coupling amount and data governance coupling edge between the logical data objects are re-determined, and the remote sensing data governance unit is re-determined based on the common remaining consumption processes, thus obtaining the reconstructed remote sensing data governance unit. Based on the available storage capacity of heterogeneous storage sources, the data capacity of each logical data object in the reconstructed remote sensing data governance unit, and the set of effective physical storage locations, the target aggregation storage source corresponding to the reconstructed remote sensing data governance unit is re-determined.

9. A remote sensing big data governance method supporting heterogeneous storage according to claim 1, characterized in that, The convergence process of the temporary collaborative replica includes: For logical data objects with temporary collaborative copies, when the corresponding set of remaining consumption processes is empty, the temporary collaborative copy is deleted, and the physical storage location corresponding to the temporary collaborative copy is deleted from the set of valid physical storage locations. When the corresponding set of remaining consumption processes is not empty and the logical data object belongs to the reconstructed remote sensing data governance unit, determine whether the physical storage location corresponding to the temporary collaborative copy corresponds to the target aggregation storage source re-determined by the remote sensing data governance unit; if they correspond, retain the temporary collaborative copy; if they do not correspond and there is already a physical storage location in the set of valid physical storage locations that corresponds to the re-determined target aggregation storage source, delete the temporary collaborative copy. When the corresponding set of remaining consumption processes is not empty and the logical data object does not belong to the reconstructed remote sensing data governance unit, and the logical data object has other valid physical storage locations, the temporary collaborative copy is deleted. Update the set of valid physical storage locations based on whether the temporary collaborative copy is retained or deleted.

10. A remote sensing big data governance system supporting heterogeneous storage, characterized in that, include: The basic relationship construction module is used to acquire heterogeneous storage sources, remote sensing production data, production processes and process input-output relationships, establish logical data objects containing spatial processing range, data capacity and effective physical storage location set, and establish consumption relationships based on process input-output relationships; The coupling relationship determination module is used to obtain the execution status of the production process, determine the set of remaining consumption processes based on the consumption relationship and the execution status, and determine the dual coupling quantity based on the spatial processing range of any two logical data objects and the set of remaining consumption processes, and establish a data governance coupling edge. The governance unit construction module is used to form candidate connected components based on the data governance coupling edges, and to split the candidate connected components according to the common remaining consumption process to obtain a set of remote sensing data governance units. The aggregation storage determination module is used to determine the target aggregation storage source for each remote sensing data governance unit based on the available storage capacity of heterogeneous storage sources, the data capacity of logical data objects, and the set of effective physical storage locations. The collaborative copy formation module is used to copy logical data objects that are not located in the target aggregation storage source to the target aggregation storage source to form temporary collaborative copies, and update the set of valid physical storage locations; The closed-loop reconstruction module is used to update the set of remaining consumption processes after the production process is completed, establish new logical data objects for the intermediate data generated by the production process, and redetermine the data governance coupling edge and remote sensing data governance unit; redetermine the target aggregation storage source according to the reconstructed remote sensing data governance unit, copy the logical data objects not located in the target aggregation storage source to the target aggregation storage source to form a new temporary collaborative copy, converge the original temporary collaborative copy, and update the set of effective physical storage locations.