An on-orbit 3D modeling method and system

Abstract

This application relates to the field of 3D modeling technology, and more particularly to an on-orbit 3D modeling method and system. The method includes: mapping user-inputted task requirements to task specification objects; performing a mixed search of a task template library and verifying the map consistency of the retrieved task templates; after passing the verification, sorting the candidate templates; expanding high-scoring candidate templates into an executable high-scoring operator sequence as the task chain output; mapping each operator in the task chain to an MCP call node to obtain the MCP call chain; calling data processing tools and 3D modeling tools to generate a preliminary 3D model, then performing incremental correction and local optimization through a dynamic feedback update mechanism, and downloading the updated incremental 3D data to a ground data center. This achieves automated task parsing without manual intervention. It possesses lightweight modeling capabilities, can perform knowledge-driven completion, and realizes a collaborative mode of rapid on-orbit response + fine ground reconstruction.

Description

Technical Field

[0001] This application relates to the field of 3D modeling technology, and in particular to an on-orbit 3D modeling method and system. Background Technology

[0002] With the rapid development of remote sensing satellite technology, the application of 3D geographic information acquisition in disaster emergency response, urban monitoring, and national defense security is becoming increasingly widespread. Currently, mainstream 3D modeling methods, such as point cloud reconstruction techniques based on motion-reconstruction structures and multi-view stereo vision, and commercial software platforms like RealityCapture and Pix4D, generally rely on ground-based computing environments for processing. These methods typically require professionals to perform complex process configurations and parameter adjustments, including multiple interactive steps such as camera parameter settings, matching point selection, and error optimization. This highly manual processing mode makes it difficult to adapt to the unattended autonomous operation requirements in an on-orbit environment.

[0003] Meanwhile, transmitting massive amounts of raw image data entirely to the ground for processing would consume huge amounts of space-to-ground communication bandwidth and ground storage resources, and introduce significant processing latency. In application scenarios with extremely high timeliness requirements, such as disaster monitoring and battlefield situational awareness, it cannot meet the real-time response requirements of acquiring local three-dimensional information within minutes. Summary of the Invention

[0004] To address the above problems, this application provides an on-orbit 3D modeling method and system, wherein the method includes the following steps:

[0005] S1. Use controlled semantic parsing to map user-inputted task requirements into task specification objects;

[0006] S2. Based on the task specification object, perform a mixed search on the task template library, and perform a graph consistency check on the retrieved task templates regarding template preconditions and data dependencies. After the check passes, sort the candidate templates.

[0007] S3. Expand the sorted high-scoring candidate templates into an executable high-scoring operator sequence using a task planning method based on a hierarchical task network and / or a planning domain description language, and output the high-scoring operator sequence as a task chain.

[0008] S4. Map each operator in the task chain to an MCP (Model Context Protocol) call node, and generate a call order based on the directed acyclic graph topology and resource availability of the task chain to obtain the MCP call chain;

[0009] S5. According to the MCP call chain, call the data processing tool and the 3D modeling tool to generate a preliminary 3D model. Then, perform incremental correction and local optimization on the multi-temporal images through a dynamic feedback update mechanism, and download the updated incremental 3D data to the ground data center.

[0010] Specifically, the structure of the task specification object includes at least the following fields: task intent, imaging modality, preconditions, and postconditions;

[0011] When the AI ​​Agent performs a mixed search of the task template library in step S2, it includes:

[0012] Based on the task specification object, a triple retrieval mechanism is used to perform parallel or serial retrieval on the task template library. The retrieval results of the triple retrieval mechanism are combined to obtain the task template. The triple retrieval mechanism includes a semantic vector index, a structured inverted index, and a task graph index. The semantic vector index uses an embedding model to vectorize the description and tags of the task template to achieve semantic similarity matching. The structured inverted index performs fast filtering on the task intent, the imaging modality, the preconditions, and the postconditions. The task graph index constructs a searchable graph of the task template according to the preconditions and postconditions to achieve subgraph matching and dependency verification.

[0013] Specifically, when the AI ​​Agent executes S2, it also executes S6 in parallel, wherein S6 includes:

[0014] Knowledge is retrieved from the extended knowledge base based on the task specification object, and the knowledge is used as a constraint to constrain the task template, the data processing tool, and the 3D modeling tool. The extended knowledge base includes at least one of urban planning rules, historical 3D models, and geographic information system data.

[0015] Specifically, S3 further includes:

[0016] When the high-resolution operator sequence is output as the task chain, a probe subtask is inserted into the high-resolution operator sequence to quickly evaluate the on-orbit environmental state parameters and drive the adaptive branch selection of the task chain based on the evaluation results.

[0017] By using constraint programming or heuristic scheduling methods, when generating the task chain, multi-objective scheduling optimization is performed with the goal of optimizing latency, downlink bandwidth, energy consumption and quality indicators to generate a near-optimal task chain that satisfies multi-objective constraints.

[0018] Specifically, the data processing tool includes multiple dedicated processing modules controlled by a modular protocol for MCP calls, the dedicated processing modules including:

[0019] MCP-SAR is used to process SAR remote sensing data and perform radiometric correction, geometric correction, and interferometric analysis.

[0020] MCP-VIS is used to process visible light images, performing cloud removal, orthorectification, and registration.

[0021] MCP-TIR is used to process thermal infrared images for nighttime modeling or to supplement surface thermal features;

[0022] MCP-3D Matching is used for multi-view geometric sparse point cloud matching.

[0023] MCP-3D Gaussian Splatting is used for efficient rendering and continuous 3D reconstruction.

[0024] MCP-Data Overlay is used for multimodal overlay and data fusion.

[0025] MCP - Data Compression, used for data compression and optimization;

[0026] The AI ​​Agent invokes the dedicated processing module according to the MCP call chain and outputs the processed standardized data to the 3D modeling tool for use in the modeling process.

[0027] Specifically, each of the dedicated processing modules declares its capabilities by providing a standardized tool capability declaration file. Each tool capability declaration file includes the schema definition of input and output, adjustable parameters and their value ranges, resource profiles, telemetry fields, error models and rollback lists, and declares the version and compatibility strategy.

[0028] The dedicated processing module supports idempotent operations, health checks, checkpoint output, and permission auditing, and provides cost indication capabilities through Dryrun mode.

[0029] The input and output schema definitions in the tool capability declaration file are used for interface alignment and automatic verification; the adjustable parameters and value ranges are used for parameter binding in the constraint planning stage; the resource profile is used for resource allocation in multi-objective scheduling optimization; the telemetry field is used for real-time monitoring of module execution status and triggering quality gating; the error model and fallback list are used to define failure fallback paths and retry strategies; and the version and compatibility strategy are used to ensure the compatibility of module calls.

[0030] The idempotent operation is achieved by hashing the input and parameters to generate an idempotent key, and repeated calls with the same input return consistent results;

[0031] The health check periodically reports the module's status indicators;

[0032] The checkpoint output supports resuming task execution from breakpoints;

[0033] The Dry run mode verifies the feasibility of inputs, parameters, resource estimates, and data paths without performing core computations.

[0034] Specifically, S1 includes: converting user natural language input into a structured task specification object through constraint hints and JSON Schema validation, the task specification object containing at least some of the fields of task intent, region, time window, imaging modality, quality target, latency target, resource constraints, privacy constraints, and expected output, and strictly mapping the user input into unambiguous structured data through controlled semantic parsing.

[0035] Specifically, when the AI ​​Agent performs the graph consistency check on the retrieved task template regarding template preconditions and data dependencies in step S2, it includes:

[0036] Check the satisfiability of the task template in terms of input data availability, tool capability matching, and dependency closure. Use the task graph index to construct a searchable graph of the template preconditions and postconditions. Verify the dependency consistency of data flow and task chain through subgraph matching.

[0037] In addition, the system of the present invention operates by the method described above, including:

[0038] The user requirements layer is used to map user-inputted task requirements into task specification objects using controlled semantic parsing.

[0039] Based on the task specification object, a hybrid search is performed on the task template library, and the retrieved task templates are checked for consistency in terms of template preconditions and data dependencies. After the check passes, the candidate templates are sorted.

[0040] The sorted high-scoring candidate templates are expanded into an executable sequence of high-scoring operators using a task planning method based on a hierarchical task network and / or a planning domain description language, and the high-scoring operator sequence is output as a task chain.

[0041] The intelligent agent scheduling layer is used to map each operator in the task chain to an MCP calling node, and generate a calling order based on the topological order of the directed acyclic graph and resource availability of the task chain to obtain the MCP calling chain.

[0042] The data processing tool and 3D modeling tool are invoked according to the MCP call chain to generate a preliminary 3D model. Then, the multi-temporal images are incrementally corrected and locally optimized through a dynamic feedback update mechanism, and the updated incremental 3D data is downloaded to the ground data center.

[0043] The toolset layer includes an extensible modular toolset, which consists of multiple dedicated processing modules.

[0044] This application has the following technical effects:

[0045] This invention introduces an AI Agent-driven automated task parsing and scheduling mechanism, enabling fully autonomous processing from user requirements to 3D model generation in an on-orbit environment, effectively overcoming the strong dependence of existing ground modeling methods on human intervention.

[0046] Through the collaboration of mission planning and modular toolchain, data screening and preliminary modeling were completed on-board, significantly reducing the bandwidth pressure and processing latency of raw data transmission.

[0047] Meanwhile, with the help of dynamic feedback and incremental update capabilities, it can respond to real-time task requirements within minutes, comprehensively improving the timeliness and practicality of on-orbit 3D modeling. Attached Figure Description

[0048] The above and other objects, features, and advantages of exemplary embodiments of this application will become readily understood by reading the following detailed description with reference to the accompanying drawings. Several embodiments of this application are illustrated in the drawings by way of example and not limitation, and the same or corresponding reference numerals denote the same or corresponding parts.

[0049] Figure 1 This is a flowchart illustrating the steps of an on-orbit 3D modeling system according to an embodiment of this application;

[0050] Figure 2 This is a schematic diagram of the framework of an on-orbit 3D modeling method in an embodiment of this application. Detailed Implementation

[0051] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0052] This invention provides a systematic solution to key problems existing in current on-orbit 3D modeling technologies. Traditional 3D modeling methods heavily rely on ground computing platforms and manual intervention, failing to meet the requirements of autonomy, real-time performance, and resource constraints in an on-orbit environment. Specifically, existing technologies face three fundamental challenges: First, the modeling process requires continuous participation from professionals in parameter adjustment and process control, which inherently conflicts with the unattended operation mode in on-orbit environments; second, the raw remote sensing data volume is enormous, and complete downloading would lead to extreme strain on communication bandwidth and a significant increase in processing latency; third, the response cycle from data acquisition to model generation is too long, making it difficult to meet the urgent needs of emergency monitoring and other scenarios requiring minute-level responses. The root cause of these problems lies in the fact that the existing technical architecture has failed to effectively integrate capabilities such as intelligent task planning, modular tool scheduling, and dynamic optimization into the on-orbit processing stage.

[0053] The core of this invention lies in the construction of a complete on-orbit 3D modeling system, such as... Figure 1 As shown, this system uses AIAgent as the intelligent scheduling hub, realizing fully autonomous processing from task reception to model generation. In the specific implementation process, it follows the principles outlined below. Figure 2 Perform the steps shown:

[0054] S1. Use controlled semantic parsing to map the user's input task requirements to a task specification object MissionSpec;

[0055] S2. Based on the task specification object MissionSpec, perform a mixed search on the task template library, and perform a graph consistency check on the retrieved task templates regarding template preconditions and data dependencies. After the check passes, sort the candidate templates.

[0056] S3. Expand the sorted high-scoring candidate templates into an executable high-scoring operator sequence using a task planning method based on hierarchical task network HTN and / or planning domain description language PDDL, and output the high-scoring operator sequence as a task chain.

[0057] S4. Map each operator in the task chain to an MCP call node, and generate the call order according to the DAG topology order and resource availability of the task chain to obtain the MCP call chain.

[0058] S5. According to the MCP call chain, call the data processing tool and 3D modeling tool to generate a preliminary 3D model. Then, through the dynamic feedback update mechanism, perform incremental correction and local optimization on the multi-temporal images, and download the updated incremental 3D data to the ground data center.

[0059] In this embodiment, the user submits modeling requests via natural language, such as a 3D reconstruction of a city. AIAgent first employs controlled semantic parsing technology, using constraint hints and JSON Schema validation to convert the user's natural language input into a structured task specification object (MissionSpec). The MissionSpec contains at least some fields from the following: task intent, region, time window, imaging modality, quality objective, latency objective, resource constraints, privacy constraints, and expected output. Controlled semantic parsing rigorously maps the user input into unambiguous structured data. Constraint hints and JSON Schema validation ensure the accuracy of the semantic mapping, avoiding ambiguity issues inherent in natural language understanding.

[0060] Based on the mission specification object MissionSpec, the system initiates a hybrid retrieval mechanism to select suitable processing flows from the mission template library. This stage employs a three-pronged retrieval strategy working in parallel. The structure of the mission specification object MissionSpec includes at least the following fields: mission intent, imaging modality, preconditions, and postconditions.

[0061] When the AI ​​Agent performs a mixed retrieval of the task template library in S2, it includes:

[0062] Based on the task specification object MissionSpec, a triple retrieval mechanism is used to search the task template library in parallel or sequentially. The retrieval results from the triple retrieval mechanism are combined to obtain the task templates. The triple retrieval mechanism includes a semantic vector index, a structured inverted index, and a task graph index. The semantic vector index uses an embedding model to vectorize the description and tags of the task templates to achieve semantic similarity matching. The structured inverted index quickly filters the task intent, imaging modality, preconditions, and postconditions. The task graph index constructs a searchable graph of the task templates according to the preconditions and postconditions to achieve subgraph matching and dependency verification. The semantic vector index is responsible for matching similar tasks at the semantic level, the structured inverted index quickly filters for explicit task attributes, and the task graph index ensures the feasibility of the task chain by analyzing the dependencies between templates. This multi-dimensional retrieval mechanism improves the accuracy and efficiency of template selection. The retrieved candidate templates also need to undergo graph consistency verification to verify their satisfies in terms of input data availability, tool capability matching, and dependency closure. Only templates that pass the verification will enter the subsequent processing stage.

[0063] In this embodiment, when the AI ​​Agent performs graph consistency verification on the retrieved task template regarding template preconditions and data dependencies in S2, it includes:

[0064] Check the satisfiability of the task template in terms of input data availability, tool capability matching, and dependency closure. Use the task graph index to construct a searchable graph of the template preconditions and postconditions. Verify the dependency consistency of data flow and task chain through subgraph matching.

[0065] The system's knowledge enhancement layer, RAG, provides domain knowledge support for the modeling process through retrieval-enhanced generation techniques. While executing S2, AIAgent also executes S6 in parallel. S6 includes:

[0066] The system retrieves knowledge from an extended knowledge base based on the task specification object MissionSpec, and uses this knowledge as constraints to govern the task template, data processing tools, and 3D modeling tools. The extended knowledge base includes at least one of urban planning rules, historical 3D models, and geographic information system (GIS) data. When processing 3D modeling tasks for urban areas, the system can retrieve relevant urban planning rules, building codes, and historical 3D models from the extended knowledge base, incorporating this knowledge into the modeling process as constraints. For example, when encountering areas with missing imagery data, the system can make reasonable inferences based on prior knowledge such as building height restrictions in urban planning, improving the completeness and rationality of the generated model.

[0067] After the task chain is generated, the system maps each operator to a corresponding MCP call node and generates a specific call order based on the DAG topology of the task chain and the current resource availability. The MCP call chain is constructed using standardized interface definitions. Each dedicated processing module clearly defines its input / output format, adjustable parameter range, resource requirements, and other key information through a tool capability declaration file (Manifest). This standardized design ensures seamless collaboration between different modules and provides good scalability for the system. During module calls, the system uses an idempotent key mechanism to ensure consistent results for repeated calls with the same input, effectively improving the system's fault tolerance and execution stability. Specifically, when using the high-resolution operator sequence as the task chain output, probe subtasks are inserted into the high-resolution operator sequence to quickly evaluate on-orbit environmental state parameters and drive adaptive branch selection of the task chain based on the evaluation results.

[0068] By using constraint programming or heuristic scheduling methods, multi-objective scheduling optimization is performed when generating task chains, with the goal of optimizing latency, downlink bandwidth, energy consumption and quality indicators, in order to generate near-optimal task chains that meet multi-objective constraints.

[0069] Specifically, each dedicated processing module declares its capabilities by providing a standardized tool capability declaration file (Manifest). Each tool capability declaration file (Manifest) includes the schema definition of input and output, adjustable parameters and their value ranges, resource profiles, telemetry fields, error models and rollback lists, and declares the version and compatibility policy.

[0070] The dedicated processing module supports idempotent operations, health checks, checkpoint output, and permission auditing, and provides cost indication capabilities through Dry run mode.

[0071] Among them, the schema definition of input and output in the tool capability declaration file Manifest is used for interface alignment and automatic verification; adjustable parameters and value ranges are used for parameter binding in the constraint planning stage; resource profiles are used for resource allocation in multi-objective scheduling optimization; telemetry fields are used for real-time monitoring of module execution status and triggering quality gating; error models and fallback lists are used to define failure fallback paths and retry strategies; and version and compatibility strategies are used to ensure the compatibility of module calls.

[0072] Idempotent operations are achieved by hashing the input and parameters to generate idempotent keys, and repeated calls with the same input return consistent results.

[0073] The module status indicators for health checks are reported regularly.

[0074] Checkpoint output supports resuming task execution from breakpoints;

[0075] Dry run mode verifies the feasibility of inputs, parameters, resource estimates, and data paths without performing core computations.

[0076] In the data processing and 3D modeling stages of this embodiment, the system is equipped with multiple dedicated processing modules to handle different types of remote sensing data. The data processing tools include multiple dedicated processing modules controlled by the MCP via a modular protocol. These dedicated processing modules include:

[0077] MCP-SAR is used to process SAR remote sensing data and perform radiometric correction, geometric correction, and interferometric analysis.

[0078] MCP-VIS is used to process visible light images, performing cloud removal, orthorectification, and registration.

[0079] MCP-TIR is used to process thermal infrared images for nighttime modeling or to supplement surface thermal features;

[0080] MCP-3D Matching is used for multi-view geometric sparse point cloud matching.

[0081] MCP-3D Gaussian Splatting is used for efficient rendering and continuous 3D reconstruction.

[0082] MCP-Data Overlay is used for multimodal overlay and data fusion.

[0083] MCP - Data Compression, used for data compression and optimization;

[0084] The AI ​​Agent invokes a dedicated processing module based on the MCP call chain and outputs the processed standardized data to the 3D modeling tool for use in the modeling process.

[0085] The MCP-SAR module specifically processes synthetic aperture radar data, performing operations such as radiometric correction, geometric correction, and interferometric analysis. The MCP-VIS module is responsible for cloud removal, orthorectification, and registration of visible light imagery. The MCP-TIR module focuses on the analysis of thermal infrared imagery, providing support for nighttime modeling or supplementing surface thermal features. These modules adopt a unified interface specification, ensuring both flexibility in invocation and consistency in data processing quality. The MCP-3D Matching module achieves sparse point cloud matching through multi-view geometry methods to construct a preliminary 3D structure; the MCP-3D Gaussian Splatting module is responsible for efficient rendering and continuous 3D reconstruction, enhancing the realism and visualization of the model. These modules were designed with full consideration of the computational resource limitations of the on-orbit environment, achieving a good balance between processing efficiency and model quality through algorithm optimization. In addition, the system also includes the MCP-Data Overlay module for multimodal data fusion and the MCP-DataCompression module for data compression and optimization, ensuring that the generated 3D model can be efficiently downloaded to the ground.

[0086] In addition, the system features a dynamic feedback update mechanism. After generating the initial 3D model, the system can perform incremental correction and local optimization based on multi-temporal imagery. This mechanism continuously monitors modeling quality indicators and automatically triggers corresponding compensation measures when model deviations or data gaps are detected. For example, if cloud cover is detected causing a decrease in local model accuracy, the system can autonomously schedule data from other temporal phases or modalities for supplementary processing, ensuring the stability of the output model's quality. Simultaneously, the incremental update mechanism ensures that only the changed parts need to be reprocessed and re-uploaded, significantly reducing data transmission volume.

[0087] Obviously, the embodiments described above are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0088] It should be understood that when the terms "first," "second," etc., are used in the claims, description, and drawings of this application, they are only used to distinguish different objects and not to describe a specific order. The terms "comprising" and "including" used in the description and claims of this application indicate the presence of the described features, integrals, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or collections thereof.

Claims

1. An on-orbit three-dimensional modeling method, executed by an AI Agent, characterized by, The AI Agent comprises the following steps when performing the method: S1, using controlled semantic parsing to map the user input task requirement into a task specification object; S2, based on the task specification object, performing hybrid retrieval on the task template library, and performing graph consistency verification on the retrieved task templates regarding template preconditions and data dependencies, and after verification, sorting the candidate templates; S3, expanding the sorted high-score candidate templates into executable high-score operator sequences through a task planning method based on hierarchical task networks and / or planning domain description language, and outputting the high-score operator sequences as task chains; S4, mapping each operator in the task chain to an MCP call node, and generating a call order according to the directed acyclic graph topology order of the task chain and resource availability to obtain an MCP call chain; S5, calling data processing tools and three-dimensional modeling tools according to the MCP call chain to generate a preliminary three-dimensional model, and then performing incremental correction and local optimization on multi-temporal images through a dynamic feedback updating mechanism, and downloading the updated incremental three-dimensional data to a ground data center.

2. The method of claim 1, wherein, The structure of the task specification object includes at least the fields of task intent, imaging modality, precondition and postcondition; When performing the hybrid retrieval on the task template library in S2, the AI Agent comprises: Based on the task specification object, using a triple retrieval mechanism to perform parallel or serial retrieval on the task template library, and obtaining the task templates by integrating the retrieval results of the triple retrieval mechanism, wherein the triple retrieval mechanism includes a semantic vector index, a structured inverted index and a task graph index, the semantic vector index performs vectorization representation on the description and label of the task template through an embedding model to realize semantic similarity matching, the structured inverted index filters the task intent, the imaging modality, the precondition and the postcondition fields, and the task graph index constructs the task templates into a searchable graph according to the preconditions and the postconditions to realize subgraph matching and dependency verification.

3. The method of claim 1, wherein, When performing S2, the AI Agent also performs S6 in parallel, and the S6 comprises: Retrieving knowledge from an extended knowledge base according to the task specification object, and using the knowledge as a constraint condition to constrain the task templates, the data processing tools and the three-dimensional modeling tools, wherein the extended knowledge base includes at least one of urban planning rules, historical three-dimensional models and geographic information system data.

4. The method of claim 1, wherein, The S3 further comprises: When outputting the high-score operator sequences as the task chains, inserting a detection subtask in the high-score operator sequences for evaluating on-orbit environmental state parameters, and driving adaptive branch selection of the task chains according to the evaluation results; Through constraint planning or heuristic scheduling methods, when generating the task chains, performing multi-objective scheduling optimization to generate a near-optimal task chain that satisfies multi-objective constraints, with the optimization of time delay, downlink bandwidth, energy consumption and quality indicators as the target.

5. The method of claim 1, wherein, The data processing tool comprises a plurality of special processing modules for invoking a modular protocol control (MCP), and the special processing modules comprise: MCP-SAR, for processing SAR remote sensing data, performing radiation correction, geometric correction, and interference analysis; MCP-VIS, for processing visible light images, completing cloud removal, orthorectification, and registration; MCP-TIR, for processing thermal infrared images, for night modeling or ground surface thermal feature supplement; MCP-3D Matching, for multi-view geometric sparse point cloud matching; MCP-3D Gaussian Splatting, for efficient rendering and continuous three-dimensional reconstruction; MCP-Data Overlay, for multi-modal overlay and data fusion; MCP-Data Compression, for data compression and optimization; The AI Agent invokes the special processing modules according to the MCP call chain, and outputs the processed standardized data to the three-dimensional modeling tool for use in the modeling process.

6. The method of claim 5, wherein, Each special processing module declares its capabilities by providing a standardized tool capability declaration file, each tool capability declaration file comprising Schema definition of input and output, adjustable parameters and value range, resource profile, telemetry field, error model and fallback list, and declaring version and compatibility strategy; The special processing module supports idempotent operation, health check, checkpoint output and permission audit, and provides cost prompt capability through Dry run mode The Schema definition of input and output in the tool capability declaration file is used for interface alignment and automatic verification, the adjustable parameters and value range are used for parameter binding in the planning stage, the resource profile is used for resource allocation in multi-target scheduling optimization, the telemetry field is used for real-time monitoring of module execution state and triggering of quality gating, the error model and fallback list are used for defining failure fallback path and retry strategy, and the version and compatibility strategy are used for ensuring compatibility of module calling; The idempotent operation is implemented by generating an idempotent key through hashing of input and parameters, and repeated calling of the same input returns consistent results; The health check regularly reports module state indicators; The checkpoint output supports breakpoint continuation of task execution; The Dry run mode verifies the feasibility of input, parameters, resource estimation and data path without performing core calculation.

7. The method of claim 1, wherein, The S1 comprises: converting user natural language input into a structured task specification object through constraint prompt and JSONSchema verification, the task specification object containing at least part of fields of task intent, area, time window, imaging modality, quality target, time delay target, resource constraint, privacy constraint and expected output, and strictly mapping the user input into unambiguous structured data through controlled semantic analysis.

8. The method of claim 1, wherein, When performing the graph consistency check on the retrieved task template in the S2 regarding template preconditions and data dependencies, the AI Agent comprises: The satisfiability of the task template is checked in terms of input data availability, tool capability matching and dependency closure, the template preconditions and postconditions are constructed into searchable graphs using the task graph index, and the dependency consistency of data flow and task chain is verified through subgraph matching.

9. An on-orbit three-dimensional modeling system operating using the method of any of claims 1-8, wherein, The method comprises the following steps: a user requirement layer is used to map the user input task requirement into a task specification object by controlled semantic parsing; a task template library is searched based on the task specification object, and the retrieved task templates are checked for graph consistency in terms of template preconditions and data dependencies, and the candidate templates are sorted after the check is passed; the high-score candidate templates after sorting are unfolded into executable high-score operator sequences by a task planning method based on hierarchical task network and / or planning domain description language, and the high-score operator sequences are output as task chains; an agent scheduling layer is used to map each operator in the task chain into an MCP calling node, and to generate a calling sequence according to the directed acyclic graph topology order of the task chain and resource availability to obtain an MCP calling chain; a data processing tool and a three-dimensional modeling tool are called according to the MCP calling chain to generate a preliminary three-dimensional model, then the multi-temporal image is incrementally corrected and locally optimized through a dynamic feedback updating mechanism, and the updated incremental three-dimensional data is transmitted to a ground data center; a tool layer comprises an extensible modular tool set composed of multiple special processing modules.

Images