A layered self-adaptive multi-modal flight data intelligent association alignment method
By processing multimodal flight data using a hierarchical adaptive method and utilizing physical constraints and cross-modal attention modules, high-precision, interpretable, and adaptive correlation of multi-source flight data is achieved. This solves the problems of heterogeneity and spatiotemporal differences in multimodal data and improves the accuracy and adaptability of the correlation.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SICHUAN UNIV
- Filing Date
- 2026-03-30
- Publication Date
- 2026-06-26
AI Technical Summary
Existing technologies struggle to effectively handle the heterogeneity and spatiotemporal differences of multimodal flight data, resulting in unreliable correlation results. They are unable to adaptively meet the different needs of real-time navigation and post-analysis, and lack a systematic processing framework with physical constraints.
A hierarchical adaptive approach is adopted, which introduces the physical dynamic model of the aircraft as a constraint, utilizes generative adversarial networks and cross-modal attention modules to map multi-source flight data to a common feature space, performs local precise association and global semantic association in parallel, and outputs elastic association results.
It improves the accuracy and robustness of multimodal flight data association, ensures that the results conform to physical laws, adapts to different mission requirements, and enhances the system's flexibility and efficiency in complex environments.
Smart Images
Figure CN122286646A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the interdisciplinary fields of avionics, flight data fusion and artificial intelligence, and more specifically, to a hierarchical adaptive intelligent association and alignment method for multimodal flight data. Background Technology
[0002] Modern aircraft carry an increasing number of sensors, such as optical cameras, infrared detectors, radar, lidar, and various navigation devices, generating massive amounts of heterogeneous flight data. This data exhibits significant differences in modality, format, spatiotemporal reference, and sampling rate, forming "data silos" that severely restrict the realization of advanced applications such as flight situational awareness, autonomous navigation, and health management.
[0003] The existing technologies have the following main shortcomings: First, traditional methods are mostly designed for data from the same or dual sources, and lack a systematic processing framework for complex correlation scenarios of multimodal data (visible light, infrared, point clouds, etc.). Second, most data-driven methods are pure end-to-end models, ignoring crucial physical prior knowledge such as aircraft kinematics and dynamics, resulting in unreliable correlation results in cases of scarce or abnormal data, and even violating common sense in physics. Third, the correlation strategies are rigid and cannot adaptively output correlation results of different granularities according to navigation tasks with high real-time requirements or post-analysis tasks with high accuracy requirements.
[0004] Therefore, there is an urgent need for a new technical solution that can integrate physical mechanisms with deep learning, adaptively process multimodal data, and output interpretable and reliable correlation results. Summary of the Invention
[0005] To address the aforementioned issues, this application provides a hierarchical adaptive intelligent association and alignment method for multimodal flight data. This method aims to overcome the deficiencies of existing technologies and offers a hierarchical, adaptive, physical-data dual-driven, and flexible output approach for multimodal flight data association and alignment. Through a hierarchical processing architecture, this method addresses data unification, feature alignment, intelligent association, and adaptive output issues step by step, significantly improving the accuracy, robustness, and practicality of multi-source flight data association.
[0006] The first aspect of this invention provides a hierarchical adaptive intelligent association and alignment method for multimodal flight data, comprising: Acquire multi-source airborne flight data and perform preprocessing; The data of different modes in the preprocessed multi-source airborne flight data are mapped to a common feature space to obtain cross-modal aligned features, wherein the physical dynamic model of the aircraft is introduced as a constraint in the mapping process; Based on the cross-modal alignment features, local precise association and global semantic association are performed in parallel; wherein, the local precise association is used to obtain pixel-level or point-level matching results, and the global semantic association is used to obtain globally consistent semantic association results; Based on the results of the local precise association and the global semantic association, association information of the corresponding granularity is obtained by fusing them.
[0007] In one optional implementation, the multi-source airborne flight data includes at least multi-source GPS / IMU signals, visible light images, infrared thermal images, synthetic aperture radar images, and lidar point clouds.
[0008] In one optional implementation, the data of different modalities in the preprocessed multi-source airborne flight data are mapped to a common feature space, and an improved generative adversarial network is used, whose generator includes a spatial adaptive normalization module. The loss function of the improved generative adversarial network is:
[0009]
[0010] in, For cycle-consistent loss, To counteract the generation loss, For the structure preservation loss term based on phase consistency, and For the source image and generating images Phase consistency diagram.
[0011] In one optional implementation, the introduction of the aircraft physical dynamics model as a constraint specifically includes: Construct a physical information neural network layer, which embeds the six-degree-of-freedom rigid body dynamics equations of the aircraft. The cross-modal alignment features are decoded into flight state parameters, which are then input into the physical information neural network layer to obtain predicted state parameters based on the recursion of the dynamic equations. The physical consistency loss is calculated based on the predicted state parameters and the measured state parameters, specifically as follows:
[0012] in, For physical consistency loss, These are the predicted state parameters for the next time step. Extract the true state parameters from the measured data at the next moment; This loss is added as a regularization term to the total training loss of the network used to map data from different modalities to a common feature space.
[0013] In one optional implementation, the execution of global semantic association specifically involves: Using the observation samples acquired by each sensor in the preprocessed multi-source airborne flight data as nodes and the similarity between the cross-modal alignment features as edges, an association graph is constructed. By solving the problem of minimizing the normalized graph cut objective function constructed based on the normalized Laplacian matrix of the association graph, the global association problem is transformed into the partitioning problem of the association graph. By solving the spectral relaxation solution of the normalized graph cut optimization problem, a global association result satisfying the cycle consistency constraint is obtained.
[0014] In one alternative implementation, after obtaining the global association results that satisfy the cycle consistency constraint, a cross-modal joint attention module is introduced; This module uses the features of the current flight mission as the query, and the cross-modal alignment features as the key and value, to calculate the importance weight of each modality of data through a multi-head attention mechanism. The cross-modal alignment features are weighted and fused according to the weights to obtain a task-adaptive unified association representation as the result of the global semantic association.
[0015] In one alternative implementation, it further includes: Monitor the distribution differences between newly acquired multi-source airborne flight data and the training data used to train the mapping process and the global semantic association process; When the distribution difference exceeds a preset threshold, a lightweight fine-tuning is triggered on the improved generative adversarial network used in the mapping process and the multi-head attention network used for weighted fusion in the global semantic association process.
[0016] A second aspect of this invention provides a hierarchical adaptive intelligent association and alignment device for multimodal flight data, the device comprising: The data preprocessing module is used to acquire and preprocess multi-source airborne flight data; The physical constraint feature alignment module is used to map data of different modes in the preprocessed multi-source airborne flight data to a common feature space to obtain cross-modal aligned features. The physical dynamic model of the aircraft is introduced as a constraint during the mapping process. A multi-granularity parallel association decision module is used to perform local precise association and global semantic association in parallel based on the cross-modal alignment features; wherein, the local precise association is used to obtain pixel-level or point-level matching results, and the global semantic association is used to obtain globally consistent semantic association results; and The task-oriented flexible output module fuses the results of the local precise association and the global semantic association to obtain association information of corresponding granularity.
[0017] A third aspect of the present invention provides an electronic device, characterized in that it includes: a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement a hierarchical adaptive intelligent association and alignment method for multimodal flight data.
[0018] A fourth aspect of this invention provides a computer-readable storage medium, characterized in that a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, a hierarchical adaptive intelligent association and alignment method for multimodal flight data is provided.
[0019] This application has at least the following advantages or beneficial effects: (1) It has high correlation accuracy and strong robustness. Through hierarchical processing and physical constraints, it effectively overcomes the challenges brought by noise, occlusion and multimodal differences, and can still maintain reliable correlation performance in complex and extreme flight environments.
[0020] (2) It has good physical interpretability, abandons the "black box" model, and integrates physical dynamics as a clear constraint into the learning process, ensuring that the correlation results conform to the basic physical laws and enhancing the credibility and acceptability of the results.
[0021] (3) The system has strong self-adaptability and its flexible output architecture can meet different needs from airborne real-time processing to ground-based in-depth analysis; the online learning mechanism enables the system to continuously evolve and adapt to changes in missions and sensors.
[0022] (4) Strong global consistency: By introducing cyclic consistency constraints based on graph theory and spectral clustering, the inherent problems of conflict and inconsistency in the association results in traditional methods are fundamentally solved, providing mathematically self-consistent and logically reliable input for subsequent multi-view data fusion, which greatly improves the accuracy and robustness of the fusion results.
[0023] (5) High task adaptability: The innovative cross-modal joint attention module enables the data association process to dynamically adjust the dependence weights on different modal information according to real-time task requirements, realizing a leap from "data-driven" to "task-driven" and significantly improving the system's flexibility and efficiency in complex dynamic environments. Attached Figure Description
[0024] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0025] Figure 1 This is a flowchart of a hierarchical adaptive intelligent association and alignment method for multimodal flight data proposed in an embodiment of this application; Figure 2 This is a structural diagram of a hierarchical adaptive multimodal flight data intelligent association and alignment device proposed in an embodiment of this application; Figure 3 This is a schematic diagram of cross-modal feature alignment proposed in an embodiment of this application; Figure 4 This is a schematic diagram of a multi-granularity intelligent association decision-making structure proposed in an embodiment of this application; Figure 5 This is a schematic diagram of an electronic device according to this application. Detailed Implementation
[0026] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0027] Please refer to Figure 1 , Figure 1 This is a flowchart of a hierarchical adaptive intelligent association and alignment method for multimodal flight data proposed in an embodiment of this application. Figure 1 As shown, a hierarchical adaptive intelligent association and alignment method for multimodal flight data includes: S100: Acquire multi-source airborne flight data and perform preprocessing; In this embodiment, the multi-source airborne flight data set is It includes at least multi-source GPS / IMU signals, visible light imagery, infrared thermal imaging, synthetic aperture radar imagery, and lidar point clouds. The acquired multi-source airborne flight data undergoes spatiotemporal benchmark unification and quality enhancement preprocessing. Timestamp sequences and spatial reference system information from each data stream are extracted. Time interpolation is used to unify all data onto the same high-precision time axis. Using the fused high-confidence GPS / INS trajectory as a benchmark, coordinate transformation is employed to unify all spatial data into the same navigation coordinate system. A spatiotemporal context completion algorithm based on tensor decomposition is used to repair missing data. Simultaneously, outlier data removal and data denoising are performed to generate a benchmark spatiotemporal trajectory and output a unified spatiotemporal dataset with enhanced quality.
[0028] S200: Map the data of different modes in the preprocessed multi-source airborne flight data to a common feature space to obtain cross-modal aligned features, wherein the physical dynamic model of the aircraft is introduced as a constraint during the mapping process; In this embodiment, for image-type data, an improved generative adversarial network is used to transform heterogeneous images into a common feature space. The improved generative adversarial network is an improved HC_CycleGAN network. The main improvements are the addition of a SPADE spatial adaptive normalization module to the generator and the addition of a phase-consistency-based structure-preserving loss to the loss function. This network preserves the geometric details of the image through the spatial adaptive normalization residual module and the phase-consistency structure-preserving loss. Figure 3 As shown; the loss function of the improved generative adversarial network is the total loss, which includes cycle consistency loss, adversarial generation loss, and a structure preservation loss term based on phase consistency, specifically:
[0029] in, For cycle-consistent loss, To counteract the generation loss, The loss term is for structure preservation based on phase consistency.
[0030] The structure-preserving loss term based on phase consistency is obtained by calculating the mean square error of the phase consistency maps of the source and generated images using a log-Gabor filter bank. The specific function is as follows:
[0031] in, and For the source image and generating images Phase consistency diagram.
[0032] For trajectory and state data, smooth motion feature vectors are extracted. During feature extraction and mapping, a physical dynamics model of the aircraft is introduced as a constraint (the physical dynamics model is a well-known six-degree-of-freedom rigid body dynamics model in the aerospace field, a general existing technology; this application only embeds it as a physical constraint into the neural network, without any improvement to the dynamics model itself). Specifically: a physical information neural network layer is constructed, embedding the six-degree-of-freedom rigid body dynamics equations of the aircraft within this layer; cross-modal aligned features are decoded into flight state parameters, input into the physical information neural network layer, and the predicted state parameters for the next moment are obtained through numerical integration recursion; the mean square error between the actual state parameters for the next moment extracted from the measured data and the predicted state parameters is calculated to obtain the physical consistency loss; this loss is added as a regularization term to the total training loss of the network used to map data from different modalities to a common feature space, penalizing feature expressions that violate the laws of physical motion.
[0033] S300: Based on the cross-modal alignment features, perform local precise association and global semantic association in parallel; wherein, the local precise association is used to obtain pixel-level or point-level matching results, and the global semantic association is used to obtain globally consistent semantic association results; In this embodiment, local precise association employs a dynamic adaptive rotation matching algorithm. Based on the initial matching, it estimates the local affine deformation field for elastic correction, achieving sub-pixel / sub-point level alignment and outputting a precise transformation matrix, such as... Figure 4 As shown. The global semantic association specifically involves: using the observation samples acquired by each sensor in the preprocessed multi-source airborne flight data as nodes, and the similarity between cross-modal alignment features as edges, constructing an association graph; transforming the global association problem into a normalized graph cut optimization problem of the association graph, whose objective function is constructed based on the normalized Laplacian matrix of the association graph, and its expression is:
[0034]
[0035] in, It is a relationship diagram The normalized Laplace matrix, It is the Laplace matrix. It is a variant of the degree matrix. It is the total number of observations. This represents the number of real entities to be estimated. This is achieved by using a normalized Laplace matrix instead of the standard Laplace matrix.
[0036] By solving the spectral relaxation solution of the normalized graph cut optimization problem, the row vectors of the solution are clustered to obtain a global association result that satisfies the cycle consistency constraint. After obtaining the global association result that satisfies the cycle consistency constraint, a cross-modal joint attention module is introduced. This module uses the features of the current flight mission as the query and the cross-modal alignment features as the key and value. It calculates the importance weight of each modality's data through a multi-head attention mechanism, and its expression is:
[0037]
[0038] in, For query, As key, The value is...
[0039] The cross-modal alignment features are weighted and fused according to their weights to obtain a task-adaptive unified association representation as the final global semantic association result.
[0040] S400: Based on the results of the local precise association and the global semantic association, association information of the corresponding granularity is obtained by fusing them.
[0041] In this embodiment, the system maintains a "task-output granularity" mapping table. Based on the different downstream task requirements such as real-time navigation, post-event analysis, and real-time obstacle avoidance, it flexibly outputs association information of different granularities, such as offset matrices for local precise association and association reports for global semantic association. At the same time, it monitors the distribution difference between newly acquired multi-source airborne flight data and training data. When the distribution difference exceeds a preset threshold, it triggers lightweight fine-tuning of the improved generative adversarial network and the multi-head attention network in the cross-modal joint attention module, and continuously optimizes the association model through an online learning mechanism.
[0042] Please refer to Figure 2 , Figure 2 This is a structural diagram of a hierarchical adaptive multimodal flight data intelligent association and alignment device proposed in an embodiment of this application. Figure 2 As shown in the figure, this disclosure also provides a hierarchical adaptive multimodal flight data intelligent association and alignment device, the device comprising: a data preprocessing module 201, a physical constraint feature alignment module 202, a multi-granularity parallel association decision module 203, and a mission-oriented elastic output module 204; wherein, Data preprocessing module 201 is used to acquire multi-source airborne flight data and perform preprocessing. The physical constraint feature alignment module 202 is used to map the data of different modes in the preprocessed multi-source airborne flight data to a common feature space to obtain cross-modal alignment features, wherein the aircraft physical dynamic model is introduced as a constraint during the mapping process; The multi-granularity parallel association decision module 203 is used to perform local precise association and global semantic association in parallel based on the cross-modal alignment features; wherein, the local precise association is used to obtain pixel-level or point-level matching results, and the global semantic association is used to obtain globally consistent semantic association results; and The task-oriented flexible output module 204 fuses the results of the local precise association and the global semantic association to obtain association information of corresponding granularity.
[0043] This disclosure also provides an electronic device, please refer to... Figure 5 , Figure 5 This is a schematic diagram of an electronic device illustrated in an embodiment of this disclosure. For example... Figure 5 As shown, the electronic device 100 includes a memory 110 and a processor 120. The memory 110 and the processor 120 are connected via a bus for communication. The memory 110 stores a computer program that can run on the processor 120 to implement the steps in the hierarchical adaptive multimodal flight data intelligent association and alignment method disclosed in this embodiment.
[0044] The disclosed embodiments also provide a computer-readable storage medium that, when the instructions in the computer-readable storage medium are executed by a processor of a computer device, enables the computer device to perform steps in the hierarchical adaptive multimodal flight data intelligent association and alignment method as described in the embodiments of this disclosure.
[0045] This application describes embodiments with reference to flowchart illustrations and / or block diagrams of methods, apparatuses, electronic devices, and computer program products according to embodiments of this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0046] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0047] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0048] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.
[0049] Although preferred embodiments of the present application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present application.
[0050] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.
[0051] The above provides a detailed description of the hierarchical adaptive intelligent association and alignment method for multimodal flight data provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the method and its core ideas. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.
Claims
1. A hierarchical adaptive intelligent association and alignment method for multimodal flight data, characterized in that, include: Acquire multi-source airborne flight data and perform preprocessing; The data of different modes in the preprocessed multi-source airborne flight data are mapped to a common feature space to obtain cross-modal aligned features, wherein the physical dynamic model of the aircraft is introduced as a constraint in the mapping process; Based on the cross-modal alignment features, local precise association and global semantic association are performed in parallel; wherein, the local precise association is used to obtain pixel-level or point-level matching results, and the global semantic association is used to obtain globally consistent semantic association results; Based on the results of the local precise association and the global semantic association, association information of the corresponding granularity is obtained by fusing them.
2. The hierarchical adaptive intelligent association and alignment method for multimodal flight data according to claim 1, characterized in that, The multi-source airborne flight data includes at least multi-source GPS / IMU signals, visible light images, infrared thermal images, synthetic aperture radar images, and lidar point clouds.
3. The hierarchical adaptive intelligent association and alignment method for multimodal flight data according to claim 1, characterized in that, The data from different modalities in the preprocessed multi-source airborne flight data are mapped to a common feature space. An improved generative adversarial network (GAN) is used, whose generator includes a spatial adaptive normalization module. The loss function of the improved GAN is: in, For cycle-consistent loss, To counteract the generation loss, For the structure preservation loss term based on phase consistency, and For the source image and generating images Phase consistency diagram.
4. The hierarchical adaptive intelligent association and alignment method for multimodal flight data according to claim 1, characterized in that, The specific constraints introduced by the aircraft physical dynamics model include: Construct a physical information neural network layer, which embeds the six-degree-of-freedom rigid body dynamics equations of the aircraft. The cross-modal alignment features are decoded into flight state parameters, which are then input into the physical information neural network layer to obtain predicted state parameters based on the recursion of the dynamic equations. The physical consistency loss is calculated based on the predicted state parameters and the measured state parameters, specifically as follows: in, For physical consistency loss, These are the predicted state parameters for the next time step. Extract the true state parameters from the measured data at the next moment; This loss is added as a regularization term to the total training loss of the network used to map data from different modalities to a common feature space.
5. The hierarchical adaptive intelligent association and alignment method for multimodal flight data according to claim 1, characterized in that, The execution of global semantic association specifically refers to: Using the observation samples acquired by each sensor in the preprocessed multi-source airborne flight data as nodes and the similarity between the cross-modal alignment features as edges, an association graph is constructed. By solving the problem of minimizing the normalized graph cut objective function constructed based on the normalized Laplacian matrix of the association graph, the global association problem is transformed into the partitioning problem of the association graph. By solving the spectral relaxation solution of the normalized graph cut optimization problem, a global association result satisfying the cycle consistency constraint is obtained.
6. The hierarchical adaptive intelligent association and alignment method for multimodal flight data according to claim 5, characterized in that, After obtaining the global association results that satisfy the cycle consistency constraint, a cross-modal joint attention module is introduced; This module uses the features of the current flight mission as the query, and the cross-modal alignment features as the key and value, to calculate the importance weight of each modality of data through a multi-head attention mechanism. The cross-modal alignment features are weighted and fused according to the weights to obtain a task-adaptive unified association representation as the result of the global semantic association.
7. The hierarchical adaptive intelligent association and alignment method for multimodal flight data according to claim 1, characterized in that, Also includes: Monitor the distribution differences between newly acquired multi-source airborne flight data and the training data used to train the mapping process and the global semantic association process; When the distribution difference exceeds a preset threshold, a lightweight fine-tuning is triggered on the improved generative adversarial network used in the mapping process and the multi-head attention network used for weighted fusion in the global semantic association process.
8. A hierarchical adaptive intelligent association and alignment device for multimodal flight data, characterized in that, The device includes: The data preprocessing module is used to acquire and preprocess multi-source airborne flight data; The physical constraint feature alignment module is used to map data of different modes in the preprocessed multi-source airborne flight data to a common feature space to obtain cross-modal aligned features. The physical dynamic model of the aircraft is introduced as a constraint during the mapping process. A multi-granularity parallel association decision module is used to perform local precise association and global semantic association in parallel based on the cross-modal alignment features; wherein, the local precise association is used to obtain pixel-level or point-level matching results, and the global semantic association is used to obtain globally consistent semantic association results; and The task-oriented flexible output module fuses the results of the local precise association and the global semantic association to obtain association information of corresponding granularity.
9. An electronic device, characterized in that, include: A processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the hierarchical adaptive multimodal flight data intelligent association alignment method according to any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium, which, when executed by a processor, implements the hierarchical adaptive multimodal flight data intelligent association and alignment method as described in any one of claims 1 to 7.