Remote sensing service integrated service system

Abstract

The invention provides a remote sensing service integrated service system, which comprises an algorithm service platform, a data service platform, a task management platform and a unified monitoring platform, and is characterized in that the algorithm service platform is used for providing operator registration management and business process arrangement service, and comprises an operator management module for realizing operator registration, operator test and operator service release; the scene arrangement module comprises a manual arrangement unit and an AI arrangement unit and is used for realizing scene arrangement, scene test and scene service release; the data service platform is used for realizing unified storage and data service release of multi-source heterogeneous data; the task management platform is used for performing scheduling, real-time monitoring and visual management on tasks; and the unified monitoring platform provides cross-platform cluster resource monitoring and management services. Through centralized management of various operators, business processes and data, unified algorithm, data and service support can be provided for various business systems, so that the development and management efficiency of the business systems is improved, and the cost is reduced.

Description

Technical Field

[0001] The present invention relates to the field of cloud computing technology, and in particular to a remote sensing business integrated service system. Background Art

[0002] The application of cloud computing technology in remote sensing data processing has become the core driving force for industry transformation. Through flexible resource scheduling, distributed storage and parallel computing capabilities, it has significantly improved the processing efficiency and analysis depth of massive remote sensing data.

[0003] Remote sensing image processing systems are diverse. In existing technologies, these systems are typically developed, used, and operated independently, with corresponding data resources managed separately. However, across different remote sensing image processing systems, many business scenarios utilize all or some of the same algorithms, and their data resource warehousing and storage requirements are also consistent. Developing and operating multiple systems independently results in repeated algorithm development and data resource warehousing, increasing duplication of effort and unnecessary algorithm development and data maintenance costs. Therefore, there is an urgent need to explore a solution that can provide unified algorithm, data, and service management for multiple business systems. Summary of the Invention

[0004] Provides a comprehensive remote sensing business service system that can provide unified algorithms, data and service support for multiple business systems, thereby improving the efficiency of algorithm and data utilization, and improving the development and maintenance efficiency and reducing costs of various business systems.

[0005] A remote sensing business integrated service system is provided, characterized in that the system includes an algorithm service platform, a data service platform, a task management platform and a unified monitoring platform, wherein:

[0006] The algorithm service platform is used to provide operator registration management and business process orchestration services, including:

[0007] The operator management module is used to implement operator registration, operator testing, operator service release, and operator lifecycle management;

[0008] Scenario orchestration module, used to implement scenario orchestration, scenario testing, and scenario service release, including manual orchestration units and AI orchestration units;

[0009] The data service platform is used to realize the unified storage of multi-source heterogeneous data and the release of data services;

[0010] The task management platform is used to schedule, monitor and visually manage tasks in real time;

[0011] The unified monitoring platform provides cross-platform cluster resource monitoring and management services.

[0012] Furthermore, the manual arrangement unit is used to construct a visual business process diagram by dragging operator nodes and linking the operator nodes in a logical order.

[0013] The AI choreography unit is used to automatically construct a business process diagram corresponding to the target scenario based on the scenario choreography text instructions carrying the target scenario information.

[0014] Furthermore, the AI orchestration unit includes an instruction receiving window, an orchestration processing subunit and a scene display interface, wherein:

[0015] The instruction receiving window is used to receive a scene arrangement task instruction in text format input by a user, wherein the scene arrangement task instruction carries the target scene information required for arranging the target scene;

[0016] The orchestration processing subunit is used to determine one or more target operators corresponding to the target scenario and link sequence information between the target operators based on the target scenario information, and to connect the target operators in series according to the link sequence to construct a business process diagram corresponding to the target scenario;

[0017] The scenario display interface is used to display the business process diagram.

[0018] Furthermore, the scene arrangement module provides a scene service API interface, and the scene service is called by the business system through the scene service API interface.

[0019] Furthermore, the data service platform includes:

[0020] Data source configuration module, used to configure library table data source and file data source;

[0021] The data warehousing module is used to establish collection tasks to automatically store and analyze multi-source heterogeneous data and automatically publish the data services.

[0022] Furthermore, the data warehousing module provides a data service API interface, and the data service is called by the business system through the data service API interface.

[0023] Furthermore, the data service platform supports the storage of data types including satellite image data, radar data, aerial survey data, terrain data and vector data.

[0024] Furthermore, the operator management module includes:

[0025] The basic image repository is used to provide standardized running container images in different language environments;

[0026] The operator package verification unit is used to verify whether the operator package file meets the algorithm development specifications.

[0027] Furthermore, the task management platform includes:

[0028] The scheduling configuration module is used for scheduling configuration management and provides services for configuring scheduling policies on demand;

[0029] The task scheduling module is used to dynamically allocate computing resources based on the real-time resource consumption of tasks, support the mixed deployment of deep learning tasks and conventional tasks, and support concurrent task scheduling;

[0030] The task monitoring and management module is used to monitor and display the real-time resource consumption and running status of tasks, and supports manual adjustment of task sequence.

[0031] The beneficial effects of the present invention include at least:

[0032] 1. This invention centrally manages multiple operators, business processes, and data, providing unified algorithms, data, and service support for multiple business systems, thereby improving the development and management efficiency and reducing costs of each business system.

[0033] 2. This invention integrates multiple operators through the algorithm service platform and provides a manual scenario arrangement service. By dragging and combining operators, scenarios can be quickly arranged, which is conducive to the rapid implementation of customized application business process design for different business types and improves the flexibility of scenario arrangement.

[0034] 3. This invention integrates multiple operators through the use of an algorithm service platform to provide AI automatic scenario arrangement services. By allowing users to input text instructions including target scenario information, a business process diagram corresponding to the target scenario can be automatically generated, allowing non-professionals to use the system, thereby improving the system's intelligence and ease of use.

[0035] 4. The present invention realizes automatic data storage and automatic data service release through the data service platform, provides data services when the business system calls scenario services, so that one piece of data can be used for multiple business systems, improves data utilization efficiency, and reduces data management difficulty. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0037] Figure 1This is a schematic diagram of the overall architecture of the remote sensing business integrated service system of the present invention;

[0038] Figure 2 This is a schematic diagram of the algorithm service platform structure of the present invention;

[0039] Figure 3 This is a schematic diagram of the AI orchestration unit structure of the present invention;

[0040] Figure 4 This is a structural diagram of the data service platform of the present invention. DETAILED DESCRIPTION

[0041] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field are within the scope of protection of the present invention.

[0042] The Remote Sensing Integrated Service System, based on a private remote sensing cloud, provides a service-centric management and governance platform for the remote sensing cloud product ecosystem. This cloud product ecosystem includes products such as an image preprocessing system, an image coordination system, and an online interpretation and training system. The Remote Sensing Integrated Service System provides algorithmic, data, and service support for these systems.

[0043] The relevant concepts in this invention are defined as follows:

[0044] Operator: An independently running algorithmic program that implements a specific function. It can be developed using languages such as Spark, Java, Python, and Golang.

[0045] Operator package: A zip file that packages the algorithm program and algorithm description file according to certain specifications.

[0046] Operator registration: Register the algorithm package file to the algorithm service platform, where you can debug and use the algorithm.

[0047] Scenario: Generate customized business processes by dragging algorithm nodes, connecting them in a certain order and logic, and configuring algorithm parameters.

[0048] Scenario service: Publish the tested scenario as a scenario service. You can create a task for the scenario service by calling the scenario service API interface.

[0049] Data service: Publish the successfully stored data as a data service. You can create a task for the data service by calling the data service API interface.

[0050] Figure 1The overall architecture diagram of a remote sensing business integrated service system provided by the present invention is shown. The system includes an algorithm service platform, a data service platform, a task management platform and a unified monitoring platform, wherein:

[0051] The algorithm service platform is used to provide operator registration management and business process orchestration services, including:

[0052] The operator management module is used to implement operator registration, operator testing, operator service release, and operator lifecycle management;

[0053] Scenario orchestration module, used to implement scenario orchestration, scenario testing, and scenario service release, including manual orchestration units and AI orchestration units;

[0054] The data service platform is used to realize the unified storage of multi-source heterogeneous data and the release of data services;

[0055] The task management platform is used to schedule, monitor and visually manage tasks in real time;

[0056] The unified monitoring platform provides cross-platform cluster resource monitoring and management services.

[0057] The remote sensing business integrated service system is based on the remote sensing private cloud and provides a service-centric management and governance platform for the remote sensing cloud product system. Figure 1 As shown in Figure 1, the remote sensing cloud product system includes products such as the remote sensing image preprocessing system, the remote sensing image overall management system, the online interpretation and training system, and other remote sensing image processing systems. The remote sensing business integrated service system is used to provide algorithm, data, and service support for the above systems.

[0058] Based on an embodiment of the present invention, the algorithm service platform provides operator registration management and business process orchestration services, and implements the integration and processing of various remote sensing business processes based on container technology. Figure 2 The schematic diagram of the algorithm service platform is shown in FIG. Figure 2 As shown, the algorithm service platform includes an operator management module, a scenario orchestration module, an operator service management module, and a scenario service management module.

[0059] According to one embodiment of the present invention, the operator management module is used to register the operator package provided by the operator provider, test the successfully registered operators, launch the operators that pass the test, and perform life cycle management on the operators.

[0060] Specifically, operator providers encapsulate their developed algorithm programs into container images, package them into operator packages, and upload them to the operator management platform. The operator management module provides an API for registering and querying operators, which operator package providers use to register operators with the operator management module. Operator providers can upload operator packages through the API or upload them directly.

[0061] The operator management module includes a basic image repository and an operator package verification unit, wherein the basic image repository is used to provide standardized running containers in different language environments; the operator package verification unit is used to verify whether the operator package file meets the algorithm development specifications.

[0062] Operator packages must comply with operator development specifications. Before registering an operator, the image repository must store the base images that the operator can rely on. During operator registration, the operator management module verifies the compressed package hierarchy and metadata file conformance and matches the base images that meet the requirements. Successfully registered operators will be displayed in the operator management module's list, and the operator package will be stored on the server. After registration, the operator management module manages the operator's metadata information, including the operator name, operator category, operator provider, and the operator storage address, which is the access address of the operator package on the server.

[0063] Furthermore, the operator management module is used to test successfully registered operators and bring those that pass the test online. These operators are then used as component units in the scenario orchestration module for scenario orchestration. Before testing, test data files must be uploaded to the Remote Sensing Business Integrated Service System, and test parameters must be set, including the file's input and output paths.

[0064] The operator management module also manages operators throughout their lifecycle, including editing, replacing, and deleting operators, importing and exporting operators, batch importing and exporting operators, and searching for operators based on various search criteria. This export and import functionality allows exported operator packages to be imported into other environments and re-imported within the same environment after adjustments and modifications. This operator import and export facilitates operator migration and facilitates rapid deployment of the Remote Sensing Integrated Service Platform on the client side.

[0065] The operator management module is used to register and manage various operators used in various business processes in various business systems in the remote sensing cloud product system, providing business and service support for the above systems.

[0066] According to one embodiment of the present invention, a scenario orchestration module is used to implement scenario orchestration, scenario testing, and scenario service release, and includes a manual orchestration unit and an AI orchestration unit. The manual orchestration unit is used to construct a visual business process diagram by dragging operator nodes and linking them in a logical order, supporting parameter mapping and process logic configuration; the AI orchestration unit is used to automatically construct a business process diagram corresponding to the target scenario based on scenario orchestration text instructions that carry target scenario information.

[0067] Specifically, the manual orchestration unit is used to create a blank scenario canvas. By dragging operator nodes, linking each operator node in a certain order and logic, and configuring operator parameters, a visual business process diagram is constructed to achieve zero-coding and rapid generation of customized scenarios.

[0068] The business process is Figure 1 A directed graph represents the topology of a business process, with nodes representing operators and lines representing the order in which operators are executed. Operator parameters include process input parameters and hidden parameters. Process input parameters are passed in when creating a task, while hidden parameters require parameter values to be set.

[0069] According to one embodiment of the present invention, the AI orchestration unit includes an instruction receiving window, an orchestration processing subunit and a scene display interface. Figure 3 A schematic diagram of the AI orchestration unit structure is shown.

[0070] The instruction receiving window is used to receive a scene arrangement task instruction in text format input by a user, wherein the scene arrangement task instruction carries target scene information required for arranging the target scene. The target scene information refers to attribute information of the target scene, including information such as the scene name, scene category, and scene function.

[0071] The orchestration processing subunit is used to determine multiple target operators corresponding to the target scenario and the link sequence information between each target operator based on the scenario orchestration task instruction, and to connect the multiple target operators in series according to the link sequence to construct a business process diagram corresponding to the target scenario.

[0072] Continue to see Figure 3, the orchestration processing subunit includes a query subunit, a matching subunit and a construction subunit. The query subunit is used to use the target scene information to query the scene-operator relationship table, and obtain multiple target operator information corresponding to the multiple target operators required to orchestrate the target scene and the link order between each target operator. It should be noted that the scene-operator relationship table is pre-established and stored in the system by the developer, which records the mapping relationship between the scene and the multiple operators to be used to orchestrate the scene, as well as the link order information between the multiple operators. The target operator information refers to the attribute information of the target operator, including information such as operator ID, operator name, operator category and operator function.

[0073] The matching subunit is used to vectorize the target operator information to obtain the corresponding target operator vectors; use each target operator vector to search for the operator vector that matches the target operator vector from the operator vector library to obtain each matching operator vector, and determine the operator corresponding to each matching operator vector based on the operator-operator vector mapping relationship to obtain multiple matching operators, and multiple matching operators are used as the target operators required to arrange the target scene. It should be noted that the operator management module is also used to vectorize the operator information of each operator that has been successfully registered to obtain the corresponding operator vector, establish an operator-operator vector mapping relationship, and store the operator vector in the system's operator vector library. The operator information includes information such as operator ID, operator name, operator category, and operator function.

[0074] The construction subunit is used to automatically obtain the multiple target operators and determine the connection order of each target operator node according to the link order information, and construct a business process diagram corresponding to the target scenario.

[0075] The scenario display interface is used to display the business process diagram corresponding to the target scenario.

[0076] Optionally, the AI orchestration unit may further include a parameter setting subunit, configured to set parameters of each operator in the business process graph constructed by the AI orchestration unit.

[0077] Optionally, the scenario orchestration module may further include an AI orchestration adjustment unit, configured to modify the business process diagram constructed by the AI orchestration unit by manually dragging operator nodes.

[0078] According to one embodiment of the present invention, the scenario orchestration module is further used to test the created scenarios, launch the scenarios that pass the test, and automatically publish the launched scenarios as scenario services for external business systems to call.

[0079] The scene orchestration module provides a scene service API interface. The published scene service can be called by an external business system through the API interface. The business system is a remote sensing cloud product embodiment, for example, it can be a remote sensing image preprocessing system, a remote sensing image overall management system, an online training and interpretation system, and other remote sensing image processing systems.

[0080] Specifically, the business system can call published scenario services through an API interface, use the task management platform to execute the business process corresponding to the called scenario service, and retrieve the corresponding data service from the data service platform to obtain the operation results, and obtain the operation results through the API interface. The steps for the task management platform to execute the business process corresponding to the scenario include: parsing the directed graph corresponding to the scenario's business process, obtaining each operator node in the graph, loading each operator package through the server, and executing the operator package. During the execution process, the data services provided by the data management platform are used.

[0081] The scenario orchestration module also allows you to customize the parameters used to call a scenario service before it's released. Advanced scenario configuration allows you to customize the input and output parameters, simplifying parameter transmission and invocation. These customized input and output parameters are displayed in the service API description. When you call a scenario service, the system passes the customized parameters to the operator using the configured mappings, and returns the customized parameter information in the output.

[0082] The scene orchestration module is also used for the full life cycle management of scenes, including scene editing, scene copying, scene deletion, scene import and export, batch import and export of scenes, and scene retrieval according to multiple conditions.

[0083] Based on an embodiment of the present invention, the operator service management module is used to manage operator services, and supports displaying an operator service list, retrieving operator services, and viewing operator service details. Through the operator service details, the interface information and call parameter information of the operator service can be viewed.

[0084] The scene service management module is used to display the scene service list and view the scene service details. By viewing the scene service details, the business process diagram, API interface information and call parameter information of the scene service can be obtained.

[0085] The present invention integrates multiple operators through an algorithm service platform to provide a visual scene arrangement service. By dragging the corresponding operator nodes and arranging and linking each node in a certain order and logical rules, customized business process diagrams of different business types are generated; by combining operator nodes, the application business process design can be completed, making the organization and management of the business easier. It also provides AI scene arrangement services. By inputting text instructions including target scene information, a business process diagram corresponding to the target scene can be automatically generated, so that non-professionals can also use the system, improving the intelligence and ease of use of the system. The deployment or destruction operation of the scene service can be automatically completed with one click, supporting the automated calling of business processes, and saving the cost of complicated manual processing.

[0086] According to an embodiment of the present invention, the data service platform is used to realize the unified storage of multi-source heterogeneous data and data service publishing, including a data source configuration module, a data automatic storage module, a data management module and a data service management module. Figure 4 A schematic diagram of the data service platform structure is shown.

[0087] Specifically, the data service platform can uniformly access multi-source heterogeneous data, including satellite imagery, radar data, aerial survey data, terrain data, and vector data. Satellite imagery supports raw imagery packages compressed in formats such as tar, zip, rar, and 7z, compatible with various mainstream star sources, as well as imagery in formats such as tif, tiff, img, and img+ige. Radar data supports raw imagery packages compressed in L1 SLC and zip formats, as well as L2 GTC, L3APSI, and L3A SBAS imagery packages. Aerial survey data supports POS, block data, DOM, DSM, and OSGB. Terrain data supports tif and tiff formats. Vector data supports shp, mdb, gdb files, and pg databases.

[0088] The data source configuration module is used to configure library and table data sources and file data sources, and includes the library and table data source configuration unit and the file data source configuration unit. The library and table data source configuration unit is used to create library and table data sources of PostgreSQL, Oracle, and MongoDB types, set information such as the database type, name, and address, and obtain the library and table data source after a successful connectivity test. The file data source configuration unit is used to create file data sources of FTP, Alibaba Cloud OSS, and HDFS types, set configuration information, and obtain the file data source after a successful connectivity test. Configuring library and table data sources and file data sources through the data source configuration module is used to import and save data by category.

[0089] The automatic data storage module is used to establish collection tasks for the automatic storage and parsing of multi-source, heterogeneous data, and to automatically publish data services. It includes a data collection unit and a data service publishing unit. The data collection unit is used to establish collection tasks for the automatic storage of multi-source, heterogeneous data, standardize metadata information, support automatic metadata identification, and parse the collected data before categorizing and storing it by type. The data service publishing unit is used to automatically publish data services for the stored data.

[0090] Specifically, a new collection task is created in the data collection unit. This task automatically collects data from a specified directory file or library table into the corresponding data source through the data collection interface API. This supports automatic storage of satellite imagery, radar data, aerial survey data, terrain data, and vector data. Data with parsing issues can be reparsed. The data collection unit supports real-time control of the execution of a collection task through immediate execution.

[0091] The data service publishing unit is used to automatically publish data services for various types of data stored in the database. The data service platform provides a data service API interface, and the data service can be called by the business system through the data service API interface.

[0092] For optical image data, both original and result images support automatic data service publishing after storage, one service for each scene, and the published service is a raster real-time service; for vector data, one service for each layer is supported, and vector real-time service is published; for radar data, the result images include raster results and vector results, which support automatic data service publishing, and original images do not support uploading by default; for aerial survey data, DOM and DSM data are published by raster service uploading by default, and OSGB publishes 3D services by default.

[0093] The data service publishing unit supports automatically publishing data services, including real-time and cached services. For raster data, it supports real-time and cached services. For vector data, it supports real-time and cached services, as well as real-time and cached services. Terrain services convert tif-formatted data into terrain data and generate terrain tiles for external use. 3D services support publishing OSGB tiles as 3D services to provide tile data.

[0094] According to an embodiment of the present invention, the data management module is used for unified management of stored data, and the data service management module is used for unified management of data services.

[0095] Specifically, the Data Management module displays status information for various data types, such as storage method, storage time, file size, and parsing status. It can retrieve and reparse data based on various criteria, and provide data collection logs. The Data Management module also provides local upload functionality, supporting data upload via the web, such as raster data, vector data, radar data, terrain data, and aerial survey data.

[0096] The data service management module displays a list of data services, supports viewing data service details, and allows for image preview of data services. It includes a data service details unit, which provides metadata for viewing data services, such as the service ID, service name, data range, service address, and data interface API address; a data preview unit, which automatically retrieves raster and vector data for image display through data preview; and a manual publishing unit, which manually publishes raster, vector, and terrain data to data services. A resend function is provided for data services that fail to publish.

[0097] The present invention realizes automatic data storage and data service publishing through the data service platform, provides data services when external business systems call scene services, and effectively integrates local image data resources through unified source data analysis and labeling specifications, so that one piece of data can be used for multiple business systems. External business systems only need to focus on the use of data, reducing the difficulty of data management and improving remote sensing image data processing capabilities, management capabilities and service levels.

[0098] Based on one embodiment of the present invention, the task management platform is used to schedule, monitor in real time and visually manage tasks, including: a scheduling configuration module, a task scheduling module and a task monitoring module, wherein the scheduling configuration module is used to perform scheduling configuration management and provide on-demand configuration scheduling strategies; the task scheduling module is used to dynamically allocate computing resources based on the real-time resource consumption of tasks, support the mixed deployment of deep learning tasks and conventional tasks, and support concurrent task scheduling; the task monitoring management module is used to monitor and display the real-time resource consumption and operating status of the tasks, and support manual adjustment of the task order.

[0099] Specifically, the scheduling configuration module is used to configure the scheduling strategy, support on-demand configuration of scheduling strategies and related configurations, and support multi-task concurrent scheduling.

[0100] The task scheduling module dynamically allocates computing resources, supports the mixed deployment of deep learning tasks and regular business tasks, and supports concurrent task scheduling. It employs a scheduling strategy based on resource monitoring, monitoring the real-time resource consumption of tasks over a period of time and predicting future resource requirements based on this real-time resource consumption. It then dynamically allocates computing resources based on this predicted resource demand over the next period of time. Scheduling based on resource monitoring data is highly efficient.

[0101] The task monitoring and management module is used to monitor and display the task runtime, resource consumption, operating status and priority, to view the business process diagram corresponding to each task, and to manually adjust the task sorting. Specifically, the order of each operator node is displayed through the business process diagram, and the operator operation status, input and output parameters, resource consumption data and log details can be viewed by clicking on each operator node in the diagram. The resource consumption of the task, including CPU usage and memory usage, is displayed through a line chart, which facilitates real-time monitoring of the task running status. The task monitoring and management module is also used to manually adjust the task queuing order, including adjusting the task sorting order by dragging and dropping.

[0102] According to an embodiment of the present invention, the unified monitoring interface is used to provide cross-platform cluster resource monitoring, service dependency topology visualization and distributed log tracking functions, including a monitoring large screen module and a monitoring management module.

[0103] The monitoring dashboard module monitors overall platform resources. It provides an overview of task execution, displays the number of executed and queued tasks, and displays the overall utilization and remaining status of nodes in the cluster. The monitoring management module monitors resource usage for nodes and operator pods. This unified monitoring interface provides a visual display of overall system resource usage, assisting in problem location.

[0104] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or apparatus comprising the element.

[0105] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A remote sensing business integrated service system, characterized in that: The system includes an algorithm service platform, a data service platform, a task management platform and a unified monitoring platform, wherein: The algorithm service platform is used to provide operator registration management and business process orchestration services, including: The operator management module is used to implement operator registration, operator testing, operator service release, and operator lifecycle management; Scenario orchestration module, used to implement scenario orchestration, scenario testing, and scenario service release, including manual orchestration units and AI orchestration units; The data service platform is used to realize the unified storage of multi-source heterogeneous data and the release of data services; The task management platform is used to schedule, monitor and visually manage tasks in real time; The unified monitoring platform provides cross-platform cluster resource monitoring and management services.

2. The system according to claim 1, wherein: The manual arrangement unit is used to construct a visual business process diagram by dragging operator nodes and linking the operator nodes in a logical order.

3. The system according to claim 1, wherein: The AI choreography unit is used to automatically construct a business process diagram corresponding to the target scenario based on the scenario choreography text instructions carrying the target scenario information.

4. The system according to claim 3, characterized in that The AI orchestration unit includes an instruction receiving window, an orchestration processing subunit and a scene display interface, wherein: The instruction receiving window is used to receive a scene arrangement task instruction in text format input by a user, wherein the scene arrangement task instruction carries the target scene information required for arranging the target scene; The orchestration processing subunit is used to determine one or more target operators corresponding to the target scenario and link sequence information between the target operators based on the target scenario information, and to connect the target operators in series according to the link sequence to construct a business process diagram corresponding to the target scenario; The scenario display interface is used to display the business process diagram.

5. The system according to any one of claims 1 to 4, characterized in that: The scene arrangement module provides a scene service API interface, and the scene service is called by the business system through the scene service API interface.

6. The system according to claim 1, wherein: The data service platform includes: Data source configuration module, used to configure library table data source and file data source; The data warehousing module is used to establish collection tasks to automatically store and analyze multi-source heterogeneous data and automatically publish the data services.

7. The system according to claim 1, wherein: The data storage module provides a data service API interface, and the data service is called by the business system through the data service API interface.

8. The system according to any one of claims 1, characterized in that The data service platform supports the storage of data types including satellite image data, radar data, aerial survey data, terrain data and vector data.

9. The system according to claim 1, wherein: The operator management module includes: The basic image repository is used to provide standardized running container images in different language environments; The operator package verification unit is used to verify whether the operator package file meets the algorithm development specifications.

10. The method according to claim 1, characterized in that The task management platform includes: The scheduling configuration module is used for scheduling configuration management and provides services for configuring scheduling policies on demand; The task scheduling module is used to dynamically allocate computing resources based on the real-time resource consumption of tasks, support the mixed deployment of deep learning tasks and conventional tasks, and support concurrent task scheduling; The task monitoring and management module is used to monitor and display the real-time resource consumption and running status of tasks, and supports manual adjustment of task sequence.

Images