A satellite data backup method, device, equipment and medium

Abstract

The application discloses a satellite data backup method and device, equipment and medium, and relates to the field of data backup. The method comprises the following steps: extracting fingerprints from the satellite data to be backed up by using a fingerprint extraction model to obtain target data fingerprints; determining the similarity between the target data fingerprints and each backup data fingerprint; if the maximum similarity is not greater than a similarity threshold, directly backing up the satellite data to be backed up; if the maximum similarity is greater than the similarity threshold, determining the metadata of the satellite data to be backed up based on the similarity relationship between the target data fingerprints, the satellite data to be backed up and target satellite data, and backing up the metadata; the target satellite data is the backup satellite data corresponding to the most similar data fingerprint, and the most similar data fingerprint is the backup data fingerprint corresponding to the maximum similarity. The application can realize the safe protection of satellite data, reduce the storage space and transmission bandwidth required for backup, and improve the backup efficiency of satellite data.

Description

Technical Field

[0001] This invention relates to the field of data backup, and in particular to a satellite data backup method, apparatus, device, and medium. Background Technology

[0002] With the rapid development of remote sensing satellite technology, the amount of satellite data collected and the frequency of updates are growing exponentially. High-resolution Earth observation satellites can generate several terabytes of raw data every day, which requires a secure and reliable backup mechanism to prevent accidental loss. Traditional satellite data backup methods typically employ full backup or incremental backup strategies, requiring the complete transfer of raw or incremental data to the backup storage system, which has the following technical limitations: High bandwidth requirements: Satellite data volume is enormous, and a complete backup requires a large amount of network bandwidth, making the transmission process time-consuming. Especially for ground stations in remote areas, it is difficult to complete timely backup of large-scale data under limited bandwidth conditions.

[0003] High storage costs: Traditional backup methods retain large amounts of duplicate data, resulting in wasted storage resources. Although incremental backups reduce some redundancy, they still require the transfer and storage of large amounts of data.

[0004] Low backup efficiency: Existing backup systems lack intelligent analysis capabilities and cannot identify similarities and repetitive patterns between data, resulting in unnecessary backup operations.

[0005] Insufficient security: Sensitive satellite data faces the risk of leakage during transmission, and traditional encryption methods increase computational overhead and affect backup efficiency. Summary of the Invention

[0006] In view of this, the purpose of this invention is to provide a satellite data backup method, apparatus, device, and medium that can achieve secure protection of satellite data, reduce the storage space and transmission bandwidth required for backup, and reduce backup operations on unnecessary data, thereby improving the efficiency of satellite data backup. The specific solution is as follows: Firstly, this application provides a satellite data backup method, including: Acquire satellite data to be backed up, and use a target fingerprint extraction model to extract fingerprints from the satellite data to be backed up in order to obtain the corresponding target data fingerprint; Determine the first similarity between the target data fingerprint and each backup data fingerprint in the backup fingerprint database, determine the largest first target similarity from each first similarity, and determine whether the first target similarity is greater than the first similarity threshold; If the similarity of the first target is not greater than the first similarity threshold, then the satellite data to be backed up is backed up directly. If the first target similarity is greater than the first similarity threshold, then based on the similarity relationship between the target data fingerprint, the satellite data to be backed up and the first target satellite data, the metadata of the satellite data to be backed up is determined, and the metadata of the satellite data to be backed up is backed up; wherein, the first target satellite data is the backed-up satellite data corresponding to the first most similar data fingerprint, and the first most similar data fingerprint is the backup data fingerprint corresponding to the first target similarity.

[0007] Optionally, the method further includes: The training node uses its pre-loaded historical satellite data to train the local fingerprint extraction model and obtains the target model parameters of the trained fingerprint extraction model. The target model parameters of each training node are securely aggregated by the parameter server to obtain aggregated model parameters, and the target fingerprint extraction model is generated based on the aggregated model parameters.

[0008] Optionally, the step of using a target fingerprint extraction model to extract fingerprints from the satellite data to be backed up to obtain corresponding target data fingerprints includes: The convolutional neural network in the target fingerprint extraction model is used to convolve the satellite data to be backed up in order to extract the multi-scale features of the satellite data to be backed up. The multi-scale features are processed through the attention mechanism in the target fingerprint extraction model to obtain feature vectors focused on key regions, and the target data fingerprint corresponding to the satellite data to be backed up is determined based on the feature vectors. The length of the feature vector is a preset length.

[0009] Optionally, after completing the backup of the satellite data to be backed up or the metadata of the satellite data to be backed up, the method further includes: Use the target data fingerprint as the new backup data fingerprint; The new backup data fingerprint is added to the backup fingerprint database to update the backup fingerprint database.

[0010] Optionally, before determining the first similarity between the target data fingerprint and each backup data fingerprint in the backup fingerprint database, the method further includes: The second similarity between the target data fingerprint and each recent data fingerprint in the preset cache is determined, and the largest second target similarity is determined from each second similarity, and it is determined whether the second target similarity is greater than the second similarity threshold; the recent data fingerprint in the preset cache is the data fingerprint corresponding to the satellite data that was backed up within a preset time period adjacent to the current time; If the second target similarity is not greater than the second similarity threshold, then the step of determining the first similarity between the target data fingerprint and each backup data fingerprint in the backup fingerprint database is triggered; If the second target similarity is greater than the second similarity threshold, then based on the similarity relationship between the target data fingerprint, the satellite data to be backed up and the second target satellite data, the metadata of the satellite data to be backed up is determined, and the metadata of the satellite data to be backed up is backed up; wherein, the second target satellite data is the backed-up satellite data corresponding to the second most similar data fingerprint, and the second most similar data fingerprint is the recent data fingerprint corresponding to the second target similarity.

[0011] Optionally, determining the first similarity between the target data fingerprint and each backup data fingerprint in the backup fingerprint database includes: The backup engine obtains the encrypted data fingerprint sent by the fingerprint extraction component; wherein, the fingerprint extraction component is used to extract fingerprints from the satellite data to be backed up; the encrypted data fingerprint is the data fingerprint obtained by encrypting the target data fingerprint; The encrypted data fingerprint is decrypted using the backup engine to obtain the corresponding target data fingerprint, and the first similarity between the target data fingerprint and each backup data fingerprint in the backup fingerprint database is determined.

[0012] Optionally, the first similarity threshold is a value that is dynamically adjusted based on the data type and importance of the satellite data to be backed up.

[0013] Secondly, this application provides a satellite data backup device, comprising: The fingerprint extraction module is used to acquire satellite data to be backed up, and to extract fingerprints from the satellite data to be backed up using a target fingerprint extraction model to obtain the corresponding target data fingerprint. The similarity determination module is used to determine the first similarity between the target data fingerprint and each backup data fingerprint in the backup fingerprint database, and to determine the largest first target similarity from each first similarity, and to determine whether the first target similarity is greater than the first similarity threshold. The satellite data backup module is used to directly back up the satellite data to be backed up if the similarity of the first target is not greater than the first similarity threshold. The metadata backup module is used to determine the metadata of the satellite data to be backed up based on the similarity relationship between the target data fingerprint, the satellite data to be backed up, and the first target satellite data if the first target similarity is greater than the first similarity threshold, and to back up the metadata of the satellite data to be backed up; wherein, the first target satellite data is the backed-up satellite data corresponding to the first most similar data fingerprint, and the first most similar data fingerprint is the backup data fingerprint corresponding to the first target similarity.

[0014] Thirdly, this application provides an electronic device, comprising: Memory, used to store computer programs; A processor for executing the computer program to implement the aforementioned satellite data backup method.

[0015] Fourthly, this application provides a computer-readable storage medium for storing a computer program that, when executed by a processor, implements the aforementioned satellite data backup method.

[0016] In this application, satellite data to be backed up is acquired, and a target fingerprint extraction model is used to extract fingerprints from the satellite data to be backed up to obtain corresponding target data fingerprints. A first similarity is determined between the target data fingerprint and each backup data fingerprint in the backup fingerprint database, and the largest first target similarity is determined from each first similarity. It is also determined whether the first target similarity is greater than a first similarity threshold. If the first target similarity is not greater than the first similarity threshold, the satellite data to be backed up is directly backed up. If the first target similarity is greater than the first similarity threshold, the metadata of the satellite data to be backed up is determined based on the similarity relationship between the target data fingerprint, the satellite data to be backed up, and the first target satellite data, and the metadata of the satellite data to be backed up is backed up. Wherein, the first target satellite data is the backed-up satellite data corresponding to the first most similar data fingerprint, and the first most similar data fingerprint is the backup data fingerprint corresponding to the first target similarity.

[0017] Therefore, this application utilizes a target fingerprint extraction model to extract fingerprints from the satellite data to be backed up, obtaining corresponding target data fingerprints. It then calculates the similarity between the target data fingerprints and the backup data fingerprints in the backup fingerprint database. Based on this fingerprint similarity, it achieves differentiated backup of the satellite data to be backed up. Compared to directly calculating and matching similarity data directly on the satellite data, this application uses the data fingerprints corresponding to the satellite data, which can achieve secure protection of the satellite data and reduce the risk of data leakage. Furthermore, when the first target similarity is greater than a first similarity threshold, this application determines the metadata of the satellite data to be backed up based on the similarity relationship between the target data fingerprint, the satellite data to be backed up, and the first target satellite data, and backs up the metadata of the satellite data to be backed up. Compared to backing up the full or incremental satellite data, this application, by backing up the metadata of the satellite data, can significantly reduce the storage space and transmission bandwidth required for backup, and can reduce backup operations on unnecessary data, thereby improving the efficiency of satellite data backup. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0019] Figure 1 A flowchart of a satellite data backup method provided in this application embodiment; Figure 2 A backup system architecture diagram provided for an embodiment of this application; Figure 3 A flowchart of a satellite data backup process is provided for an embodiment of this application; Figure 4 A schematic diagram of a satellite data backup management page provided in an embodiment of this application; Figure 5 A schematic diagram of a satellite data backup device provided in this application embodiment; Figure 6 This is a structural diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Traditional satellite data backup methods typically employ full or incremental backup strategies, requiring the complete transfer of original or incremental data to the backup storage system. This approach suffers from several limitations: high bandwidth requirements, high storage costs, low backup efficiency, and insufficient security. To address these limitations, this application provides a satellite data backup method that achieves secure protection of satellite data, reduces the required storage space and bandwidth, minimizes backup operations on unnecessary data, and improves satellite data backup efficiency.

[0022] See Figure 1 As shown, this embodiment of the invention discloses a satellite data backup method, including: Step S11: Obtain the satellite data to be backed up, and use the target fingerprint extraction model to extract the fingerprint of the satellite data to be backed up to obtain the corresponding target data fingerprint.

[0023] In this embodiment, the satellite data to be backed up is first obtained from the data source (such as a satellite ground station), and a pre-trained target fingerprint extraction model is used to extract the fingerprint of the satellite data to be backed up, so as to obtain the target data fingerprint corresponding to the satellite data to be backed up. It should be noted that different satellite data correspond to different data fingerprints, that is, each satellite data has a unique data fingerprint, and each data fingerprint only requires a few hundred bytes to a few thousand bytes.

[0024] The training process for the target fingerprint extraction model can specifically include: training the local fingerprint extraction model using pre-loaded historical satellite data by the training nodes and obtaining the target model parameters of the trained fingerprint extraction model; securely aggregating the target model parameters of each training node through the parameter server to obtain aggregated model parameters, and generating the target fingerprint extraction model based on the aggregated model parameters.

[0025] According to one example, this embodiment first constructs a distributed training cluster on a cloud platform using a federated learning architecture. The distributed training cluster consists of multiple training nodes, each of which pre-loads a portion of historical satellite data (covering types such as multispectral, hyperspectral, and SAR (Synthetic Aperture Radar)) to ensure the generalization capability of the target fingerprint extraction model. Furthermore, the cloud platform also includes a parameter server. Specifically, the parameter server distributes the current model parameters to each training node, enabling each node to train its local fingerprint extraction model using its pre-loaded historical satellite data and the current model parameters. This results in a trained fingerprint extraction model, and the target model parameters for that model are obtained. Each training node then periodically synchronizes its target model parameters to the parameter server. Correspondingly, the parameter server securely aggregates the target model parameters from each training node to obtain aggregated model parameters. The current fingerprint extraction model is then generated based on these aggregated parameters. If the current fingerprint extraction model does not meet the preset training termination conditions, the aggregated model parameters are used as the new current model parameters, and the process returns to the previous step of distributing the current model parameters to each training node via the parameter server. If the current fingerprint extraction model meets the preset training termination conditions, it is used as the target fingerprint extraction model. These preset training termination conditions include, but are not limited to, the current fingerprint extraction model's accuracy being no less than a preset accuracy, and the current fingerprint extraction model's cumulative iteration count reaching a preset number.

[0026] It should be noted that, through the collaborative training mechanism of training nodes and parameter server, the embodiments of this application enable each training node to train the local fingerprint extraction model in parallel and perform periodic parameter synchronization, while the parameter server can use a secure aggregation algorithm to integrate model parameters and generate a global fingerprint extraction model. This not only improves model training efficiency and speeds up model training progress, but also ensures the generalization ability of the model.

[0027] Furthermore, the embodiments of this application also have a model optimization and update mechanism. Specifically, by adopting incremental learning technology, the model is fine-tuned periodically with new data to adapt to changes in data features. At the same time, the model performance monitoring module continuously evaluates the model accuracy and triggers necessary retraining.

[0028] The process of extracting fingerprints from satellite data to be backed up using a target fingerprint extraction model to obtain the corresponding target data fingerprint can specifically include: using a convolutional neural network in the target fingerprint extraction model to convolve the satellite data to be backed up in order to extract multi-scale features of the satellite data to be backed up; processing the multi-scale features through the attention mechanism in the target fingerprint extraction model to obtain feature vectors focused on key areas, and determining the target data fingerprint corresponding to the satellite data to be backed up based on the feature vectors; wherein, the vector length of the feature vector is a preset length.

[0029] Step S12: Determine the first similarity between the target data fingerprint and each backup data fingerprint in the backup fingerprint database, determine the largest first target similarity from each first similarity, and determine whether the first target similarity is greater than the first similarity threshold.

[0030] In this embodiment, after extracting the target data fingerprint corresponding to the satellite data to be backed up, a first similarity is calculated between the target data fingerprint and each backup data fingerprint in the backup fingerprint database using cosine similarity or Hamming distance. The largest first target similarity is then determined from the first similarity values, and it is then determined whether the first target similarity is greater than a first similarity threshold. The first similarity threshold is a dynamically adjusted value based on the data type and importance of the satellite data to be backed up. In addition to data type and importance, it can also be dynamically adjusted based on transmission bandwidth; for example, when bandwidth is sufficient, the similarity threshold can be lowered to preserve more detailed data.

[0031] According to one example, determining the first similarity between the target data fingerprint and each backup data fingerprint in the backup fingerprint database can specifically include: obtaining the encrypted data fingerprint sent by the fingerprint extraction component through the backup engine; wherein the fingerprint extraction component is used to extract fingerprints from the satellite data to be backed up; the encrypted data fingerprint is the data fingerprint obtained by encrypting the target data fingerprint; decrypting the encrypted data fingerprint through the backup engine to obtain the corresponding target data fingerprint, and determining the first similarity between the target data fingerprint and each backup data fingerprint in the backup fingerprint database.

[0032] In other words, after acquiring the satellite data to be backed up, the data source uses a locally embedded fingerprint extraction component to extract fingerprints from the satellite data using a target fingerprint extraction model to obtain the corresponding target data fingerprint. Then, homomorphic encryption is used to encrypt the target data fingerprint, resulting in an encrypted data fingerprint. This encrypted data fingerprint is then transmitted to the backup engine through a secure channel. Correspondingly, the backup engine receives the encrypted data fingerprint sent by the fingerprint extraction component, decrypts it to obtain the corresponding target data fingerprint, and then determines the first similarity between the target data fingerprint and each backup data fingerprint in the backup fingerprint database. It should be noted that the backup engine can use an approximate nearest neighbor algorithm to accelerate the search for large-scale backup data fingerprints in the backup fingerprint database.

[0033] In this way, the embodiments of this application transmit encrypted data fingerprints and calculate fingerprint similarity to perform differentiated backup of satellite data. Compared with directly transmitting satellite data and calculating the similarity of satellite data, the amount of data transmission can be reduced by more than 70%, which is suitable for satellite ground station environments with limited transmission bandwidth. It can also reduce the risk of satellite data leakage during transmission and achieve secure protection of satellite data.

[0034] Furthermore, before determining the first similarity between the target data fingerprint and each backup data fingerprint in the backup fingerprint database, the method further includes: determining the second similarity between the target data fingerprint and each recent data fingerprint in the preset cache, determining the largest second target similarity from each second similarity, and determining whether the second target similarity is greater than the second similarity threshold; wherein, the recent data fingerprint in the preset cache is the data fingerprint corresponding to satellite data that was backed up within a preset time period adjacent to the current time; if the second target similarity is not greater than the second similarity threshold, the step of determining the first similarity between the target data fingerprint and each backup data fingerprint in the backup fingerprint database is triggered; if the second target similarity is greater than the second similarity threshold, the metadata of the satellite data to be backed up is determined based on the similarity relationship between the target data fingerprint, the satellite data to be backed up, and the second target satellite data, and the metadata of the satellite data to be backed up is backed up; wherein, the second target satellite data is the backed-up satellite data corresponding to the second most similar data fingerprint, and the second most similar data fingerprint is the recent data fingerprint corresponding to the second target similarity.

[0035] In other words, after extracting the target data fingerprint corresponding to the satellite data to be backed up, considering that satellite data in a recent period may have high similarity, a second similarity can be calculated between the target data fingerprint and each recent data fingerprint in the preset cache. If the maximum similarity in the second similarity is greater than the second similarity threshold, then there is no need to perform similarity calculation on the large-scale backup data fingerprints in the backup fingerprint database, reducing the amount of similarity calculation and improving the backup efficiency of satellite data to a certain extent. If the maximum similarity in the second similarity is not greater than the second similarity threshold, then further similarity calculation is performed on the large-scale backup data fingerprints in the backup fingerprint database to determine the backup method for the satellite data to be backed up (backing up the satellite data or the satellite data metadata).

[0036] Step S13: If the similarity of the first target is not greater than the first similarity threshold, then the satellite data to be backed up is backed up directly.

[0037] In this embodiment of the application, if the similarity of the first target is not greater than the first similarity threshold, the satellite data to be backed up is compressed and encrypted before being uploaded to the backup storage system for backup. The compression algorithm can be LZ4 or Zstandard; the encryption algorithm can be AES-256 (Advanced Encryption Standard).

[0038] Step S14: If the first target similarity is greater than the first similarity threshold, then based on the similarity relationship between the target data fingerprint, the satellite data to be backed up and the first target satellite data, determine the metadata of the satellite data to be backed up, and back up the metadata of the satellite data to be backed up; wherein, the first target satellite data is the backed-up satellite data corresponding to the first most similar data fingerprint, and the first most similar data fingerprint is the backup data fingerprint corresponding to the first target similarity.

[0039] In this embodiment, if the first target similarity is greater than the first similarity threshold, it indicates that the satellite data to be backed up has a high similarity to the first target satellite data. In this case, direct backup of the satellite data to be backed up can be avoided. Instead, based on the data source, data descriptor, target data fingerprint, similarity relationship between the satellite data to be backed up and the first target satellite data, and the current timestamp, the metadata of the satellite data to be backed up is determined and uploaded to the backup storage system for backup. The metadata is at most 50 bytes, which significantly reduces the storage space and transmission bandwidth required for backup compared to TB-level satellite data. Furthermore, the backup process is completed in seconds. Moreover, this intelligent deduplication technology avoids duplicate storage of similar data, improving storage space utilization by more than 50% and reducing storage costs.

[0040] After backing up the metadata of the satellite data to be backed up, if you want to reconstruct the satellite data, you can determine the first target satellite data that is most similar to the satellite data based on the similarity relationship in the metadata. Then, using the data descriptor, target data fingerprint, similarity relationship and other information in the metadata, and based on the most similar first target satellite data, the satellite data can be reconstructed.

[0041] Furthermore, after completing the backup of the satellite data to be backed up or the metadata of the satellite data to be backed up, this application embodiment can also use the target data fingerprint as a new backup data fingerprint and add the new backup data fingerprint to the backup fingerprint database to dynamically update the backup fingerprint database.

[0042] It should be noted that the embodiments of this application may also employ blockchain technology to record backup operation logs, ensuring the immutability and traceability of satellite data backups. Furthermore, the embodiments of this application also feature cross-center consistency, which can be achieved by using a consistent hashing algorithm to allocate storage locations for the data to be backed up (satellite data or satellite data metadata), ensuring consistency among multiple replicas in a distributed backup environment.

[0043] Furthermore, embodiments of this application can utilize multi-core CPUs (Central Processing Units) and GPUs (Graphics Processing Units) to accelerate the fingerprint extraction and matching process, supporting data pipeline processing. Moreover, embodiments of this application also feature an incremental index update mechanism, employing an LSM tree structure (LogStructured Merge Tree) to manage the fingerprint index, achieving efficient querying and incremental updates.

[0044] Therefore, this application utilizes a target fingerprint extraction model to extract fingerprints from the satellite data to be backed up, obtaining corresponding target data fingerprints. It then calculates the similarity between the target data fingerprints and the backup data fingerprints in the backup fingerprint database. Based on this fingerprint similarity, it achieves differentiated backup of the satellite data to be backed up. Compared to directly calculating and matching similarity data directly on the satellite data, this application uses the data fingerprints corresponding to the satellite data, which can achieve secure protection of the satellite data and reduce the risk of data leakage. Furthermore, when the first target similarity is greater than a first similarity threshold, this application determines the metadata of the satellite data to be backed up based on the similarity relationship between the target data fingerprint, the satellite data to be backed up, and the first target satellite data, and backs up the metadata of the satellite data to be backed up. Compared to backing up the full or incremental satellite data, this application, by backing up the metadata of the satellite data, can significantly reduce the storage space and transmission bandwidth required for backup, and can reduce backup operations on unnecessary data, thereby improving the efficiency of satellite data backup.

[0045] See Figure 2 and Figure 3 As shown, Figure 2 The backup system architecture for implementing the satellite data backup method proposed in the embodiments of this application mainly includes four parts: cloud platform, data source (such as satellite ground station), backup engine and backup storage system. Figure 3 This document describes the specific implementation flow of the satellite data backup method proposed in the embodiments of this application.

[0046] The cloud platform includes a distributed training cluster and a parameter server. The distributed training cluster consists of multiple training nodes that work collaboratively through the parameter server. Specifically, the parameter server distributes the current model parameters to each training node, enabling each node to train its local fingerprint extraction model using its pre-loaded historical satellite data and the current model parameters. The trained node then obtains the target model parameters for the fingerprint extraction model. Each training node periodically synchronizes its target model parameters to the parameter server. Correspondingly, the parameter server securely aggregates the target model parameters from each training node to obtain aggregated model parameters. Based on these aggregated parameters, it generates the current fingerprint extraction model. If the current fingerprint extraction model does not meet the preset training termination conditions, the aggregated model parameters are used as the new current model parameters and redistributed to each training node for local fingerprint extraction model training. If the current fingerprint extraction model meets the preset training termination conditions, it is used as the target fingerprint extraction model.

[0047] The data source includes a pre-embedded fingerprint extraction component and a secure channel. Specifically, the data source acquires the satellite data to be backed up and uses the locally embedded fingerprint extraction component to extract fingerprints from the satellite data using a target fingerprint extraction model to obtain the corresponding target data fingerprint. Then, homomorphic encryption technology is used to encrypt the target data fingerprint to obtain the encrypted data fingerprint, which is then transmitted to the backup engine through the secure channel.

[0048] The backup engine includes a backup fingerprint database and a fingerprint matching processor. Specifically, the fingerprint matching processor acquires the encrypted data fingerprint sent by the fingerprint extraction component, decrypts the encrypted data fingerprint to obtain the corresponding target data fingerprint, and then determines the first similarity between the target data fingerprint and each backup data fingerprint in the backup fingerprint database. The processor then determines the largest first target similarity from the first similarity scores. If the first target similarity is not greater than a first similarity threshold, the satellite data to be backed up is compressed and encrypted before being uploaded to the backup storage system for distributed backup storage. If the first target similarity is greater than the first similarity threshold, the backup data fingerprint corresponding to the first target similarity is determined as the first most similar data fingerprint, and the backed-up satellite data corresponding to the first most similar data fingerprint is determined as the first target satellite data. Then, based on the similarity relationship between the target data fingerprint, the satellite data to be backed up, and the first target satellite data, the metadata of the satellite data to be backed up is determined, and the metadata of the satellite data to be backed up is uploaded to the backup storage system for distributed backup storage.

[0049] It should be noted that blockchain technology is used to record backup operation logs throughout the entire satellite data backup process to ensure the immutability and traceability of the satellite data backup. Furthermore, the entire satellite data backup process is supported by a security monitoring module and a performance optimization module. The security monitoring model primarily issues alerts or performs maintenance when faults or security issues occur, while the performance optimization module is mainly used to optimize and update the target fingerprint extraction model, as well as dynamically adjust similarity thresholds.

[0050] It should also be noted that the backup system used to implement the satellite data backup method proposed in the embodiments of this application mainly includes hardware implementation and software implementation.

[0051] The hardware implementation mainly consists of three parts: 1. Distributed training cluster: built using multiple high-performance GPU servers and interconnected via a high-speed InfiniBand network. The number of training nodes can be dynamically adjusted according to the data scale; 2. Fingerprint extraction component in the data source: using embedded devices or lightweight servers deployed in the data source (such as a satellite ground station), equipped with a dedicated AI (Artificial Intelligence) acceleration chip, which can provide efficient inference capabilities; 3. Backup storage system: using a distributed architecture to build a multi-node storage cluster, equipped with large-capacity hard drives and SSD (Solid State Drive) cache, supporting horizontal scaling, and each storage node is connected via dedicated lines to ensure transmission stability.

[0052] The software implementation mainly includes the training of the fingerprint extraction model, the fingerprint extraction algorithm, the similarity matching algorithm, and the intelligent backup strategy.

[0053] The fingerprint extraction model is trained using distributed training with federated learning frameworks such as TensorFlow Federated or PySyft. The initial fingerprint extraction model is based on a pre-trained ResNet-50 (Residual Network) architecture, which is fine-tuned and optimized for the characteristics of satellite data. The training process uses an asynchronous stochastic gradient descent algorithm, and after each round of training iteration, each training node uploads the model parameters to the parameter server for secure aggregation.

[0054] The fingerprint extraction algorithm uses optimization frameworks such as OpenVINO (Open Visual Inference & Neural Network Optimization, used to optimize the inference performance of deep learning models and accelerate the deployment of computer vision applications) or TensorRT (Tensor Runtime, a high-performance deep learning inference toolkit) during the model inference stage to convert the trained target fingerprint extraction model into a lightweight format and deploy it on the data source. The fingerprint extraction of the target fingerprint extraction model is based on deep feature extraction. After inputting satellite data, it is propagated forward through the network to extract the feature vector before the fully connected layer as the fingerprint.

[0055] The similarity matching algorithm uses Faiss (Facebook AI Similarity Search, an open-source similarity search tool) or similar vector search tools to build a high-performance backup engine for fingerprint similarity matching; for large-scale backup fingerprint databases, product quantization and hierarchical navigation small world (HNSW) algorithms are used to accelerate fingerprint similarity matching.

[0056] The intelligent backup strategy sets the first similarity threshold to 0.85-0.95 (dynamically adjustable). When the first target similarity is not greater than the first similarity threshold, the satellite data to be backed up is compressed, encrypted, and uploaded to the backup storage system for distributed backup storage. When the first target similarity is greater than the first similarity threshold, the metadata of the satellite data to be backed up is uploaded to the backup storage system for distributed backup storage.

[0057] In addition, such as Figure 4 As shown, the backup system also provides a visual satellite data backup management page, which allows administrators to comprehensively monitor the backup system's operating status, adjust similarity thresholds, and view backup resource saving statistics and system performance indicators (such as storage efficiency, real-time bandwidth, alarm information, backup data status, backup success rate, etc.).

[0058] In this way, the embodiments of this application have the following significant advantages over the prior art: 1. Significantly reduced bandwidth requirements: Transmitting data fingerprints of satellite data instead of satellite data reduces data transmission volume by more than 70%, adapting to the bandwidth-constrained satellite ground station environment; 2. Improved storage efficiency: Intelligent deduplication technology avoids duplicate storage of similar data, increasing storage space utilization by more than 50% and reducing storage costs; 3. Improved backup efficiency: Distributed fingerprint matching and parallel processing mechanisms improve backup processing speed by 3-5 times, supporting large-scale real-time backup of satellite data; 4. Enhanced backup security: Homomorphic encryption and blockchain technology ensure the security of fingerprint data and backup operations, preventing data leakage and tampering; 5. Good scalability: The distributed storage architecture supports elastic scaling, and the federated learning framework allows new training nodes to participate in the training of the fingerprint extraction model, adapting to the ever-growing scale of satellite data.

[0059] See Figure 5 As shown, an embodiment of the present invention discloses a satellite data backup device, comprising: The fingerprint extraction module 11 is used to acquire satellite data to be backed up, and to extract fingerprints from the satellite data to be backed up using a target fingerprint extraction model to obtain the corresponding target data fingerprint. The similarity determination module 12 is used to determine the first similarity between the target data fingerprint and each backup data fingerprint in the backup fingerprint database, and to determine the largest first target similarity from each first similarity, and to determine whether the first target similarity is greater than the first similarity threshold. The satellite data backup module 13 is used to directly back up the satellite data to be backed up if the similarity of the first target is not greater than the first similarity threshold. Metadata backup module 14 is used to determine the metadata of the satellite data to be backed up based on the similarity relationship between the target data fingerprint, the satellite data to be backed up and the first target satellite data if the first target similarity is greater than the first similarity threshold, and to back up the metadata of the satellite data to be backed up; wherein, the first target satellite data is the backed-up satellite data corresponding to the first most similar data fingerprint, and the first most similar data fingerprint is the backup data fingerprint corresponding to the first target similarity.

[0060] Therefore, this application utilizes a target fingerprint extraction model to extract fingerprints from the satellite data to be backed up, obtaining corresponding target data fingerprints. It then calculates the similarity between the target data fingerprints and the backup data fingerprints in the backup fingerprint database. Based on this fingerprint similarity, it achieves differentiated backup of the satellite data to be backed up. Compared to directly calculating and matching similarity data directly on the satellite data, this application uses the data fingerprints corresponding to the satellite data, which can achieve secure protection of the satellite data and reduce the risk of data leakage. Furthermore, when the first target similarity is greater than a first similarity threshold, this application determines the metadata of the satellite data to be backed up based on the similarity relationship between the target data fingerprint, the satellite data to be backed up, and the first target satellite data, and backs up the metadata of the satellite data to be backed up. Compared to backing up the full or incremental satellite data, this application, by backing up the metadata of the satellite data, can significantly reduce the storage space and transmission bandwidth required for backup, and can reduce backup operations on unnecessary data, thereby improving the efficiency of satellite data backup.

[0061] In some specific embodiments, the satellite data backup device further includes: The model training unit is used to train the local fingerprint extraction model using its own pre-loaded historical satellite data through the training node, and to obtain the target model parameters of the trained fingerprint extraction model. The model parameter aggregation unit is used to securely aggregate the target model parameters of each training node through the parameter server to obtain aggregated model parameters, and generate the target fingerprint extraction model based on the aggregated model parameters.

[0062] In some specific embodiments, the fingerprint extraction module 11 includes: The feature extraction unit is used to convolve the satellite data to be backed up using the convolutional neural network in the target fingerprint extraction model in order to extract the multi-scale features of the satellite data to be backed up. The fingerprint determination unit is used to process the multi-scale features through the attention mechanism in the target fingerprint extraction model to obtain a feature vector focused on the key region, and to determine the target data fingerprint corresponding to the satellite data to be backed up based on the feature vector. The length of the feature vector is a preset length.

[0063] In some specific embodiments, the satellite data backup device further includes: The fingerprint database update unit is used to use the target data fingerprint as a new backup data fingerprint and add the new backup data fingerprint to the backup fingerprint database to update the backup fingerprint database.

[0064] In some specific embodiments, the satellite data backup device further includes: The similarity determination unit is used to determine the second similarity between the target data fingerprint and each recent data fingerprint in the preset cache, and to determine the largest second target similarity from each second similarity, and to determine whether the second target similarity is greater than the second similarity threshold; the recent data fingerprint in the preset cache is the data fingerprint corresponding to satellite data that was backed up within a preset time period adjacent to the current time; The step triggering unit is used to trigger the step of determining the first similarity between the target data fingerprint and each backup data fingerprint in the backup fingerprint database if the second target similarity is not greater than the second similarity threshold. The metadata backup unit is configured to determine the metadata of the satellite data to be backed up based on the similarity relationship between the target data fingerprint, the satellite data to be backed up, and the second target satellite data if the second target similarity is greater than the second similarity threshold, and to back up the metadata of the satellite data to be backed up; wherein the second target satellite data is the backed-up satellite data corresponding to the second most similar data fingerprint, and the second most similar data fingerprint is the recent data fingerprint corresponding to the second target similarity.

[0065] In some specific embodiments, the similarity determination module 12 includes: A fingerprint acquisition unit is used to acquire encrypted data fingerprints sent by a fingerprint extraction component through a backup engine; wherein, the fingerprint extraction component is used to extract fingerprints from the satellite data to be backed up; and the encrypted data fingerprint is a data fingerprint obtained by encrypting the target data fingerprint. The similarity determination unit is used to decrypt the encrypted data fingerprint through the backup engine to obtain the corresponding target data fingerprint, and to determine the first similarity between the target data fingerprint and each backup data fingerprint in the backup fingerprint database.

[0066] Furthermore, embodiments of this application also disclose an electronic device, Figure 6 This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content of the diagram should not be construed as limiting the scope of this application.

[0067] Figure 6 This is a schematic diagram of the structure of an electronic device 20 provided in an embodiment of this application. Specifically, the electronic device 20 may include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 stores a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the satellite data backup method disclosed in any of the foregoing embodiments. Alternatively, the electronic device 20 in this embodiment may specifically be an electronic computer.

[0068] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.

[0069] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or optical disk, etc. The resources stored thereon can include operating system 221, computer program 222, etc., and the storage method can be temporary storage or permanent storage.

[0070] The operating system 221 is used to manage and control the various hardware devices on the electronic device 20 and the computer program 222, which may be Windows Server, Netware, Unix, Linux, etc. In addition to including a computer program capable of performing the satellite data backup method executed by the electronic device 20 as disclosed in any of the foregoing embodiments, the computer program 222 may further include computer programs capable of performing other specific tasks.

[0071] Furthermore, this application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned disclosed satellite data backup method. Specific steps of this method can be found in the corresponding content disclosed in the foregoing embodiments, and will not be repeated here.

[0072] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0073] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0074] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0075] 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 apparatus 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 apparatus. 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 apparatus that includes said element.

[0076] The technical solutions provided in this application have been described in detail above. 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 methods and core ideas of this application. 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 satellite data backup method, characterized in that, include: Acquire satellite data to be backed up, and use a target fingerprint extraction model to extract fingerprints from the satellite data to be backed up in order to obtain the corresponding target data fingerprint; Determine the first similarity between the target data fingerprint and each backup data fingerprint in the backup fingerprint database, determine the largest first target similarity from each first similarity, and determine whether the first target similarity is greater than the first similarity threshold; If the similarity of the first target is not greater than the first similarity threshold, then the satellite data to be backed up is backed up directly. If the first target similarity is greater than the first similarity threshold, then based on the similarity relationship between the target data fingerprint, the satellite data to be backed up and the first target satellite data, the metadata of the satellite data to be backed up is determined, and the metadata of the satellite data to be backed up is backed up; wherein, the first target satellite data is the backed-up satellite data corresponding to the first most similar data fingerprint, and the first most similar data fingerprint is the backup data fingerprint corresponding to the first target similarity.

2. The satellite data backup method according to claim 1, characterized in that, Also includes: The training node uses its pre-loaded historical satellite data to train the local fingerprint extraction model and obtains the target model parameters of the trained fingerprint extraction model. The target model parameters of each training node are securely aggregated by the parameter server to obtain aggregated model parameters, and the target fingerprint extraction model is generated based on the aggregated model parameters.

3. The satellite data backup method according to claim 1, characterized in that, The step of extracting fingerprints from the satellite data to be backed up using a target fingerprint extraction model to obtain the corresponding target data fingerprint includes: The convolutional neural network in the target fingerprint extraction model is used to convolve the satellite data to be backed up in order to extract the multi-scale features of the satellite data to be backed up. The multi-scale features are processed through the attention mechanism in the target fingerprint extraction model to obtain feature vectors focused on key regions, and the target data fingerprint corresponding to the satellite data to be backed up is determined based on the feature vectors. The length of the feature vector is a preset length.

4. The satellite data backup method according to claim 1, characterized in that, After completing the backup of the satellite data to be backed up or the metadata of the satellite data to be backed up, the process also includes: Use the target data fingerprint as the new backup data fingerprint; The new backup data fingerprint is added to the backup fingerprint database to update the backup fingerprint database.

5. The satellite data backup method according to claim 1, characterized in that, Before determining the first similarity between the target data fingerprint and each backup data fingerprint in the backup fingerprint database, the method further includes: The second similarity between the target data fingerprint and each recent data fingerprint in the preset cache is determined, and the largest second target similarity is determined from each second similarity, and it is determined whether the second target similarity is greater than the second similarity threshold; the recent data fingerprint in the preset cache is the data fingerprint corresponding to the satellite data that was backed up within a preset time period adjacent to the current time; If the second target similarity is not greater than the second similarity threshold, then the step of determining the first similarity between the target data fingerprint and each backup data fingerprint in the backup fingerprint database is triggered; If the second target similarity is greater than the second similarity threshold, then based on the similarity relationship between the target data fingerprint, the satellite data to be backed up and the second target satellite data, the metadata of the satellite data to be backed up is determined, and the metadata of the satellite data to be backed up is backed up; wherein, the second target satellite data is the backed-up satellite data corresponding to the second most similar data fingerprint, and the second most similar data fingerprint is the recent data fingerprint corresponding to the second target similarity.

6. The satellite data backup method according to claim 1, characterized in that, Determining the first similarity between the target data fingerprint and each backup data fingerprint in the backup fingerprint database includes: The backup engine obtains the encrypted data fingerprint sent by the fingerprint extraction component; wherein, the fingerprint extraction component is used to extract fingerprints from the satellite data to be backed up; the encrypted data fingerprint is the data fingerprint obtained by encrypting the target data fingerprint; The encrypted data fingerprint is decrypted using the backup engine to obtain the corresponding target data fingerprint, and the first similarity between the target data fingerprint and each backup data fingerprint in the backup fingerprint database is determined.

7. The satellite data backup method according to any one of claims 1 to 6, characterized in that, The first similarity threshold is a value that is dynamically adjusted based on the data type and importance of the satellite data to be backed up.

8. A satellite data backup device, characterized in that, include: The fingerprint extraction module is used to acquire satellite data to be backed up, and to extract fingerprints from the satellite data to be backed up using a target fingerprint extraction model to obtain the corresponding target data fingerprint. The similarity determination module is used to determine the first similarity between the target data fingerprint and each backup data fingerprint in the backup fingerprint database, and to determine the largest first target similarity from each first similarity, and to determine whether the first target similarity is greater than the first similarity threshold. The satellite data backup module is used to directly back up the satellite data to be backed up if the similarity of the first target is not greater than the first similarity threshold. The metadata backup module is used to determine the metadata of the satellite data to be backed up based on the similarity relationship between the target data fingerprint, the satellite data to be backed up, and the first target satellite data if the first target similarity is greater than the first similarity threshold, and to back up the metadata of the satellite data to be backed up; wherein, the first target satellite data is the backed-up satellite data corresponding to the first most similar data fingerprint, and the first most similar data fingerprint is the backup data fingerprint corresponding to the first target similarity.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the satellite data backup method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, Used to store a computer program, which, when executed by a processor, implements the satellite data backup method as described in any one of claims 1 to 7.

Images