Method and device for associating ship SAR data with AIS data, and electronic device
By constructing the topology and feature matching of ship SAR data and AIS data, and dynamically adjusting the cost matrix, the reliability and accuracy issues of SAR data and AIS data fusion under complex sea conditions were solved, achieving higher data association accuracy and robustness.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- WUHAN UNIV OF TECH
- Filing Date
- 2026-02-09
- Publication Date
- 2026-06-19
AI Technical Summary
Existing methods for fusing SAR and AIS data suffer from target omissions, false detections, and system positioning errors under complex sea conditions, resulting in insufficient reliability and accuracy of the fusion results.
By constructing ship topologies from SAR and AIS datasets, node feature vectors are extracted, feature distances are calculated, and a first-order cost matrix is constructed. The matching pair set is solved, and the cost matrix is dynamically adjusted to determine the target matching pair set, thus establishing the correlation between SAR and AIS data.
It improves the reliability and accuracy of the fusion results of ship SAR data and AIS data, especially the accuracy and robustness of data association in complex scenarios.
Smart Images
Figure CN122241246A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of shipbuilding, specifically to a method and apparatus for correlating ship SAR data with AIS data, electronic equipment, and computer-readable storage medium. Background Technology
[0002] With the development of global trade and increasingly busy maritime traffic, effective monitoring and accurate identification of ships at sea are crucial for ensuring maritime safety and safeguarding national maritime rights. Spaceborne Synthetic Aperture Radar (SAR) and Automatic Identification System (AIS) are two core technologies in the field of maritime surveillance: SAR is an active Earth observation system capable of real-time, all-weather, 24 / 7 observation and high-precision imaging of sea surface targets through clouds and fog; AIS is a maritime safety and communication system applied between ships, providing important dynamic and static information such as ship position, speed, heading, name, and MMSI code.
[0003] However, both of these technologies have significant limitations: SAR data only provides the physical outline and location of the target, lacking crucial semantic information such as speed and heading, and suffers from issues such as missed or false detections and system positioning errors. AIS's application is limited by signal coverage, and some vessels (such as small fishing boats or illegal vessels) may not have AIS equipment installed, or may maliciously shut it down or send false information, thus limiting the reliability and comprehensiveness of its application.
[0004] Therefore, combining the all-weather physical detection capabilities of SAR with the rich semantic information provided by AIS can create significant complementary advantages: SAR can independently detect targets without AIS enabled, providing physical verification of the authenticity of AIS data; while AIS can provide accurate identity and dynamic information annotations for detected targets in SAR images. The fusion of the two can effectively compensate for the information blind spots of a single data source, thereby constructing a more complete and reliable maritime situational awareness. By correlating the data from both sources, optimized utilization of information can be achieved, providing a highly reliable fusion solution.
[0005] However, existing SAR and AIS data fusion methods mostly employ a rigid matching strategy. This strategy heavily relies on the completeness of the observation data, but in actual SAR imaging, complex sea conditions can cause drastic changes in the overall geometry between targets, severely reducing the reliability and accuracy of the SAR and AIS data fusion results. Summary of the Invention
[0006] In view of this, it is necessary to provide a method, apparatus, electronic device and computer-readable storage medium for associating ship SAR data and AIS data, so as to achieve the technical effect of improving the reliability and accuracy of the fusion results of ship SAR data and AIS data.
[0007] To address the aforementioned technical problems, firstly, this application provides a method for correlating ship SAR data with AIS data, including:
[0008] SAR data of all ships within a defined area are acquired to form a SAR dataset, and AIS data of all ships within the defined area are acquired to form an AIS dataset. A first ship topology is constructed based on the SAR dataset, the first ship topology including multiple SAR ship nodes; a second ship topology is constructed based on the AIS dataset, the second ship topology including multiple AIS ship nodes. Feature extraction is performed on each of the SAR ship nodes to obtain SAR node feature vectors, and feature extraction is performed on each of the AIS ship nodes to obtain AIS node feature vectors. The cross operation is performed on the SAR node feature vector and the AIS node feature vector to obtain the feature distance between each SAR ship node and each AIS ship node, and a first-order cost matrix is constructed based on the feature distance. Solving the first-order cost matrix yields a first-order matching pair set, which is the set of matching pairs with the smallest sum of feature distances. The matching pair set includes multiple matching pairs, and each matching pair includes corresponding SAR data and AIS data. Based on the first-order matching pair set, a target matching pair set is determined, and the association between the SAR data and the AIS data is constructed according to the target matching pair set.
[0009] In one possible embodiment, determining the target matching pair set based on the first-order matching pair set includes: The sum of feature distances of the first-order matching pair set is taken as the sum of target feature distances; A dynamic coefficient is set based on the sum of the distances to the target features; The second-order cost matrix is obtained by adjusting the first-order cost matrix according to the dynamic coefficients. Solve the second-order cost matrix to obtain the second-order matching pair set, and use the second-order matching pair set as the target matching pair set.
[0010] In one possible embodiment, setting the dynamic coefficient based on the sum of the target feature distances includes: Based on formula The dynamic coefficients are calculated. in, 、 To set a constant, The dynamic coefficient is... It is the sum of the distances to the target features.
[0011] In one possible embodiment, adjusting the cost matrix according to the dynamic coefficients includes: Based on formula group ， Adjust the first-order cost matrix; in, For each element in the second-order cost matrix, For each element in the first-order cost matrix, The dynamic coefficient is... For the first The aforementioned SAR ship nodes For the first The aforementioned AIS ship nodes For the first topology The adjacent nodes of the SAR ship node. For the second topology The adjacent nodes of the aforementioned AIS ship nodes.
[0012] In one possible embodiment, constructing the first ship topology based on the SAR dataset includes: For a target SAR ship node, the spatial distance between other SAR ship nodes and the target SAR ship node is calculated based on the SAR data. The SAR ship node corresponding to the target spatial distance is taken as the adjacent node of the target SAR ship node. The target SAR ship node can be any SAR ship node, and the target spatial distance is the spatial distance that meets the set requirements. The target SAR ship node is unidirectionally connected to the adjacent node to form the first ship topology.
[0013] In one possible embodiment, constructing the association between the SAR data and the AIS data based on the target matching pair set includes: For any target matching pair, determine whether the SAR data and the AIS data in the target matching pair are uniquely matched; If the SAR data and AIS data in the target matching pair are not uniquely matched, the target matching pair is removed from the target matching pair set.
[0014] In one possible embodiment, constructing the association between the SAR data and the AIS data based on the target matching pair set includes: For any of the target matching pairs, the actual distance between the ships corresponding to the SAR data and the ships corresponding to the AIS data is calculated based on the SAR data and the AIS data in the target matching pair; If the actual distance of the vessel does not meet the set conditions, the target matching pair is removed from the target matching pair set.
[0015] Secondly, this application also provides a device for correlating ship SAR data and AIS data, comprising: The data acquisition module is used to acquire SAR data of all ships within a set area to form a SAR dataset and to acquire AIS data of all ships within the set area to form an AIS dataset. A topology construction module is used to construct a first ship topology based on the SAR dataset, the first ship topology including multiple SAR ship nodes, and to construct a second ship topology based on the AIS dataset, the second ship topology including multiple AIS ship nodes. The feature extraction module is used to extract features from each of the SAR ship nodes to obtain SAR node feature vectors, and to extract features from each of the AIS ship nodes to obtain AIS node feature vectors. The cost matrix construction module is used to perform cross operations on the SAR node feature vector and the AIS node feature vector to obtain the feature distance between each SAR ship node and each AIS ship node, and construct a first-order cost matrix based on the feature distance. The matrix solving module is used to solve the first-order cost matrix to obtain a first-order matching pair set. The first-order matching pair set is the matching pair set with the smallest sum of feature distances. The matching pair set includes multiple matching pairs, and each matching pair includes corresponding SAR data and AIS data. The data association module is used to determine a target matching pair set based on the first-order matching pair set, and to construct the association relationship between the SAR data and the AIS data according to the target matching pair set.
[0016] Thirdly, this application also provides an electronic device, including a memory and a processor, wherein, The memory is used to store programs; The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps in the method for associating ship SAR data and AIS data as described in any of the above implementations.
[0017] Fourthly, this application also provides a computer-readable storage medium for storing a computer-readable program or instruction, which, when executed by a processor, can implement the steps in the method for associating ship SAR data and AIS data as described in any of the above implementations.
[0018] The beneficial effects of this application are: Compared with related technologies, the method, apparatus, electronic device, and computer-readable storage medium for associating ship SAR data and AIS data provided in this application, for SAR data and AIS data of ships within a set area, constructs a first ship topology based on the SAR dataset and a second ship topology based on the AIS dataset. The feature distances between each SAR ship node and each AIS ship node are calculated based on the first and second ship topologies. A first-order cost matrix is constructed based on the feature distances, and a first-order matching pair set is obtained by solving the first-order cost matrix. A target matching pair set is determined based on the first-order matching pair set, and the association relationship between SAR data and AIS data is constructed based on the target matching pair set. Since the first-order matching pair set is the matching pair set with the smallest sum of feature distances, constructing the association relationship between SAR data and AIS data based on the first-order matching pair set can accurately associate SAR data and AIS data corresponding to the same ship. Furthermore, by linking adjacent ships based on the first and second ship topologies in this application, the reliability and accuracy of the ship SAR data and AIS data fusion results can be improved based on the relative positional relationships between ship groups. Attached Figure Description
[0019] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying 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.
[0020] Figure 1 This is a flowchart illustrating the method for associating ship SAR data and AIS data provided in an embodiment of this application. Figure 2 This is a schematic diagram of the process for constructing the first ship topology in the method for associating ship SAR data and AIS data provided in the embodiments of this application; Figure 3This is a schematic diagram of the topology of the first ship topology in the ship SAR data and AIS data association method provided in the embodiments of this application; Figure 4 This is a schematic diagram illustrating the process of determining the target matching pair set in the method for associating ship SAR data and AIS data provided in the embodiments of this application; Figure 5 A flowchart illustrating a method for associating ship SAR data with AIS data according to another embodiment of this application; Figure 6 This is a schematic diagram of the structure of the device for linking ship SAR data and AIS data provided in the embodiments of this application; Figure 7 This is a schematic diagram of the structure of an electronic device provided in one embodiment of this application. Detailed Implementation
[0021] 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 a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.
[0022] In the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more. "And / or" describes the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.
[0023] The terms "first," "second," etc., used in the embodiments of this application are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a technical feature defined with "first" or "second" may explicitly or implicitly include at least one of those features.
[0024] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.
[0025] This application provides a method and apparatus, electronic device and computer-readable storage medium for correlating ship SAR data and AIS data, which are described below.
[0026] Please refer to Figure 1 The method for associating ship SAR data and AIS data provided in this application includes: Step S101: Obtain SAR data of all ships within the designated area to form a SAR dataset, and obtain AIS data of all ships within the designated area to form an AIS dataset.
[0027] In this step, after setting the latitude and longitude geographical boundaries of the designated area and the target time range, the SAR data covering the area is retrieved and downloaded through a spaceborne SAR data platform, such as the European Space Agency Copernicus Data Space or the ASF Alaska Satellite Facility, according to the requirements of area, time, and resolution to form a SAR dataset. At the same time, the full amount of basic data of ships in the area is retrieved through the AIS system, and the AIS data of each ship is extracted to form an AIS dataset.
[0028] Furthermore, in this embodiment, before constructing the SAR dataset and AIS dataset, the SAR data and AIS data are also filtered to remove abnormal data such as ship length and beam being 0, longitude being outside the valid range (-180°~180°), latitude being outside the valid range (-90°~90°), and speed or heading values being obviously unreasonable (such as speed exceeding the ship's maximum speed), so as to avoid abnormal data interfering with subsequent analysis.
[0029] Step S102: Construct a first ship topology based on the SAR dataset, which includes multiple SAR ship nodes; construct a second ship topology based on the AIS dataset, which includes multiple AIS ship nodes.
[0030] Before constructing the first ship topology based on the SAR dataset and the second ship topology based on the AIS dataset in this step, a time calibration step is also included, which is to time-align the AIS data with the SAR data.
[0031] Different time calibration strategies are adopted based on the amount of AIS data for each ship. The specific time calibration steps include: For vessel A, if vessel A has consecutive AIS data before and after the SAR imaging time (i.e., the sampling time corresponding to the SAR data), meaning vessel A has two or more AIS data points, then a linear interpolation algorithm is used to interpolate the trajectory between consecutive time points to calculate its predicted AIS data at the SAR imaging time. The specific calculation formula is as follows: ; in, For SAR imaging time, and This provides two AIS data sets for ship A, one before and one after the SAR imaging time. , For the time of two AIS data points, These are the latitude and longitude coordinates of two AIS data sets.
[0032] If ship A has only one AIS data point before and after the SAR imaging time, this application uses speed and heading to estimate the predicted AIS data for the ship at the SAR imaging time. The specific calculation formula is as follows: ; in, SAR imaging time and the time when the ship sent AIS data Time difference, For ship speed, For the ship's course, The radius of the Earth is taken as 6,371,000 meters in this application. This indicates the ship's current latitude.
[0033] After completing the time calibration, this application also projects the SAR dataset and the AIS dataset to the same spatial coordinate system (such as the UTM coordinate system). The specific projection process includes: S1: The Y-axis of the SAR data is flipped vertically through coordinate transformation. The specific calculation formula is as follows: ;in, The coordinates are those of the original SAR image. and These represent the height and width of the original SAR image, respectively. The coordinates are the flipped SAR image coordinates.
[0034] S2: Transform the pixel coordinate system after unifying the coordinate system orientation to the WGS84 coordinate system. The specific calculation formula is as follows: ; in, These are latitude and longitude coordinates in the WGS84 coordinate system. and These are the smaller and larger longitude values corresponding to the bottom left and top right corners of the SAR image after it has been flipped. and These are the smaller and larger latitude values corresponding to the lower left and upper right corners of the SAR image after it has been flipped.
[0035] S3: Project SAR images and AIS data from the WGS84 coordinate system to the UTM coordinate system.
[0036] Specifically, through a zonal strategy, the Earth's surface is projected into multiple (e.g., 60) identical longitude zones, each covering a range of 6°, minimizing deformation distortion in each region. The specific calculation formula is as follows: ; Where East represents the eastern distance and North represents the northern distance. As a scaling factor, This is the eastward offset (in meters), used to avoid negative eastward offset values. The value is determined by whether the set region is in the Northern or Southern Hemisphere. When the set region is in the Northern Hemisphere, The value is 0, and when the set region is in the Southern Hemisphere, The value is 10,000,000 meters. From the equator to the current latitude Meridian arc length: ; in, The semi-major axis of the Earth is given by a value of rice; The square of the first eccentricity, , The flattening factor has a value of [value missing]. . The radius of curvature of the circle is: T, C, and A are auxiliary parameters, respectively: ， ， , Current longitude Interpolation with the central meridian, i.e. , The longitude of the central meridian of each longitude zone, , Longitude zone number .
[0037] After projecting the SAR and AIS datasets onto the same spatial coordinate system, a first ship topology is constructed based on the SAR dataset, and a second ship topology is constructed based on the AIS dataset within this spatial coordinate system. Please refer to [the relevant documentation / reference]. Figure 2 The first ship topology was constructed based on the SAR dataset, including: Step S201: Construct multiple SAR ship nodes in the spatial coordinate system, with each SAR ship node corresponding to unique SAR data.
[0038] Step S202: Based on the SAR data corresponding to the target SAR ship node, calculate the spatial distance between other SAR ship nodes and the target SAR ship node.
[0039] In this step, the target SAR ship node is any SAR ship node.
[0040] The spatial distance is specifically Euclidean distance, and the specific calculation formula is as follows:
[0041] in, and These are the coordinates of the two SAR ship nodes in the spatial coordinate system.
[0042] Step S203: The SAR ship node corresponding to the target spatial distance is taken as the adjacent node of the target SAR ship node, and the target spatial distance is the spatial distance that meets the set requirements.
[0043] In this step, the specific requirement is to select the K nodes with the smallest spatial distance. That is, for the target SAR ship node, select the K nodes with the smallest spatial distance to the target SAR ship node as the adjacent nodes of the target SAR ship node, and form the first ship topology by unidirectionally connecting the target SAR ship node and its corresponding adjacent nodes.
[0044] In this step, please refer to Figure 3 The one-way connection between the target SAR ship node and its corresponding adjacent nodes is specifically achieved by the target SAR ship node pointing out a one-way edge to each adjacent node.
[0045] It is understood that the specific steps for constructing the second ship topology based on the AIS dataset in this application are roughly the same as those for constructing the first ship topology based on the SAR dataset mentioned above. Similarly, multiple AIS ship nodes are constructed, the spatial distance between every two AIS ship nodes is calculated, and the K nodes with the smallest spatial distance values to each AIS ship node are selected as the adjacent nodes of each AIS ship node. The AIS ship nodes are unidirectionally connected to their corresponding adjacent nodes to form the second ship topology. For details, please refer to the aforementioned specific description, which will not be repeated here.
[0046] Step S103: Extract features from each SAR ship node to obtain SAR node feature vectors, and extract features from each AIS ship node to obtain AIS node feature vectors.
[0047] In this step, based on the node positional relationships in the first ship topology and the positional data of each SAR ship node, multi-dimensional feature extraction is performed on each SAR ship node to obtain the SAR node feature vector. ships iThe corresponding SAR node feature vector; based on the node positional relationships in the second ship topology and the positional data of each AIS ship node, multi-dimensional feature extraction is performed on each AIS ship node to obtain the AIS node feature vector. That is, the feature vector of the AIS node corresponding to ship j.
[0048] In this embodiment, multi-dimensional feature extraction is performed on each SAR and AIS ship node. Specifically, this includes extracting features from the node attributes corresponding to each SAR and AIS ship node to obtain multi-dimensional features. The node attributes corresponding to each SAR and AIS ship node specifically include the ship number, the ship's latitude and longitude coordinates, and the MMSI (Maritime Mobile Service Identity) corresponding to the actual parameters of the ship, such as these attributes.
[0049] Furthermore, the node attributes corresponding to each SAR and AIS ship node also include basic topological attributes such as node degree, local clustering coefficient, and eccentricity; and structural importance attributes such as betweenness centrality, proximity centrality, eigenvector centrality, and PageRank value. The feature values corresponding to each attribute are extracted to form the SAR node feature vector and the AIS node feature vector.
[0050] Step S104: Perform cross-operation on the SAR node feature vector and the AIS node feature vector to obtain the feature distance between each SAR ship node and each AIS ship node, and construct a first-order cost matrix based on the feature distance.
[0051] In this step, cross-operations are performed on the SAR node feature vector and the AIS node feature vector. Specifically, for any SAR node feature vector... Calculate the feature vector of each AIS node respectively. The feature distance between it and the target. The formula for calculating the feature distance is: ;in, For ships i The corresponding SAR node feature vector and the ship j The feature distance between the corresponding AIS node feature vectors.
[0052] The first-order cost matrix is constructed based on all feature distances as follows: ;in, The number of nodes representing SAR data, i.e. , The number of nodes representing AIS data, i.e. . To construct the first-order cost matrix The various elements.
[0053] Step S105: Solve the first-order cost matrix to obtain the first-order matching pair set. The first-order matching pair set is the set of matching pairs with the minimum sum of feature distances. The matching pair set includes multiple matching pairs, and each matching pair includes corresponding SAR data and AIS data.
[0054] In this step, the matching pair refers to the correspondence between SAR ship nodes and AIS ship nodes, for example, the matching pair This indicates that ship node i in the first ship topology corresponds to ship node j in the second ship topology, forming a matching pair. Each matching pair corresponds to a feature distance, and its corresponding ship node... i SAR data and ship nodes j The AIS data correspond to each other, and the set of matching pairs is the set of all matching pairs.
[0055] In this embodiment, solving for the first-order cost matrix is equivalent to finding the set of first-order matching pairs that minimizes the sum of feature distances for all matching pairs in the matching pair set. This is expressed by the formula: [Formula omitted for brevity]. The set of matching pairs that yields the minimum value is taken as the first-order matching pair set. .
[0056] Furthermore, this application specifically employs the Hungarian algorithm to solve the first-order cost matrix, obtaining a set of first-order matching pairs that minimizes the sum of feature distances.
[0057] Step S106: Determine the target matching pair set based on the first-order matching pair set, and construct the correlation between SAR data and AIS data based on the target matching pair set.
[0058] In this embodiment, please refer to Figure 4 The target matching pair set is determined based on the first-order matching pair set, specifically including: Step S401: Obtain the sum of feature distances of the first-order matching pair set as the sum of target feature distances.
[0059] In this step, the set of first-order matching pairs is substituted into the formula. The sum of target feature distances is calculated in the middle. .
[0060] Step S402: Set dynamic coefficients based on the sum of target feature distances.
[0061] In this step, the specific formula is used. The dynamic coefficients are calculated. in, , To set a constant, this embodiment sets... =0.8、 =0.4, For dynamic coefficients, This is the sum of the distances to the target features.
[0062] Step S403: Adjust the first-order cost matrix according to the dynamic coefficients to obtain the second-order cost matrix.
[0063] In this step, the specific formula group is used. , Adjust the first-order cost matrix; where, For each element in the second-order cost matrix, For each element in the first-order cost matrix, For dynamic coefficients, For the first SAR ship nodes, For the first AIS vessel nodes, For the first topology Adjacent nodes of a SAR ship node For the second topology Adjacent nodes of an AIS ship node.
[0064] Step S404: Solve for the second-order cost matrix to obtain the second-order matching pair set, and use the second-order matching pair set as the target matching pair set.
[0065] In this step, solving for the second-order cost matrix is also equivalent to finding the set of second-order matching pairs that minimizes the sum of feature distances for all matching pairs in the matching pair set. This can be expressed as: [Formula omitted for brevity]. The set of matching pairs that yields the minimum value is taken as the set of second-order matching pairs. , .
[0066] The first-order cost matrix is adjusted by setting a dynamic coefficient based on the sum of the distances to the target features to obtain the second-order cost matrix. Finally, the second-order cost matrix is solved to obtain the set of target matching pairs. The two-stage matching framework integrates location features and structural similarity information, which significantly improves the accuracy and robustness of data association in dense and complex scenarios.
[0067] It is understood that the foregoing is merely a specific example illustrating the determination of the target matching pair set based on the first-order matching pair set in this embodiment. In some embodiments of this application, the first-order matching pair set can be directly used as the target matching pair set, or it can continue to be based on, for example... Figure 4 The method shown adjusts the second-order cost matrix to obtain the third-order cost matrix, and solves the second-order cost matrix to obtain the target matching pair set, among other methods.
[0068] Compared with related technologies, the method for associating ship SAR data and AIS data provided in this embodiment involves constructing a first ship topology based on the SAR dataset and a second ship topology based on the AIS dataset for SAR and AIS data of ships within a defined area. The characteristic distances between each SAR ship node and each AIS ship node are calculated based on the first and second ship topologies. A first-order cost matrix is constructed based on the characteristic distances, and a first-order matching pair set is obtained by solving the first-order cost matrix. A target matching pair set is determined based on the first-order matching pair set, and the association relationship between SAR data and AIS data is constructed based on the target matching pair set. Since the first-order matching pair set is the matching pair set with the smallest sum of characteristic distances, constructing the association relationship between SAR data and AIS data based on the first-order matching pair set can accurately associate SAR data and AIS data corresponding to the same ship. Furthermore, by linking adjacent ships based on the first and second ship topologies in this application, the reliability and accuracy of the ship SAR data and AIS data fusion results can be improved based on the relative positional relationships between ship groups.
[0069] For further details, please refer to Figure 5 Another embodiment of this application provides a method for associating ship SAR data with AIS data, specifically including the following steps: Step S501: Obtain SAR data of all ships within the designated area to form a SAR dataset, and obtain AIS data of all ships within the designated area to form an AIS dataset.
[0070] Step S502: Construct a first ship topology based on the SAR dataset, which includes multiple SAR ship nodes; construct a second ship topology based on the AIS dataset, which includes multiple AIS ship nodes.
[0071] Step S503: Extract features from each SAR ship node to obtain SAR node feature vectors, and extract features from each AIS ship node to obtain AIS node feature vectors.
[0072] Step S504: Perform cross-operation on the SAR node feature vector and the AIS node feature vector to obtain the feature distance between each SAR ship node and each AIS ship node, and construct a first-order cost matrix based on the feature distance.
[0073] Step S505: Solve the first-order cost matrix to obtain the first-order matching pair set. The first-order matching pair set is the set of matching pairs with the minimum sum of feature distances. The matching pair set includes multiple matching pairs, and each matching pair includes corresponding SAR data and AIS data.
[0074] Step S506: Determine the target matching pair set based on the first-order matching pair set.
[0075] Step S507: Validate the target matching pair set.
[0076] In this embodiment, the target matching pair set is validated, specifically including: For any target matching pair, determine whether the SAR data and AIS data in the target matching pair are uniquely matched; if the SAR data and AIS data in the target matching pair are not uniquely matched, remove the target matching pair from the target matching pair set.
[0077] Specifically, determining whether the SAR data and AIS data in a target matching pair are uniquely corresponding means determining whether each SAR data corresponds to a unique AIS data and whether each AIS data corresponds to a unique SAR data. If so, the verification passes. Otherwise, if there is a SAR data corresponding to two or more AIS data, or an AIS data corresponding to two or more SAR data, then the target matching pairs that include these SAR data and / or AIS data are removed from the target matching pair set.
[0078] Furthermore, in some other embodiments of this application, the target matching pair set is validated, specifically including: For any target matching pair, calculate the actual distance between the ship corresponding to the SAR data and the ship corresponding to the AIS data based on the SAR data and AIS data in the target matching pair; if the actual distance between the ships does not meet the set conditions, remove the target matching pair from the target matching pair set.
[0079] The calculation of the actual distance between ships corresponding to SAR data and ships corresponding to AIS data specifically includes: Retrieve the latitude and longitude coordinates of the WGS-84 coordinate system stored as node attributes in step S102, that is, the latitude and longitude coordinates of the corresponding SAR data in each matching pair. Latitude and longitude coordinates of AIS data The Haversine formula is used to calculate the shortest distance between two points on Earth using latitude and longitude. The specific formula is as follows:
[0080] in, This represents the actual distance between ships corresponding to SAR data and ships corresponding to AIS data. Let be the average radius of the Earth, and take . rice, and These represent the difference between the latitude and longitude coordinates of the two points.
[0081] By comparing the actual distance of the ships With dynamic geographic distance threshold Remove the actual distance of the ship Greater than the dynamic geographic distance threshold The matching pairs.
[0082] Among them, determining the dynamic geographic distance threshold Specifically, this includes: calculating the geographic distance of all matched pairs that have passed the bidirectional consistency test, and obtaining the mean geographic distance. and standard deviation And based on this, according to the formula: Calculate the dynamic geographic distance threshold .
[0083] Step S508: Construct the correlation between SAR data and AIS data based on the target matching set.
[0084] It is understood that steps S501 to S506 and S508 in this embodiment are largely the same as steps S101 to S106 in the previous embodiment. For details, please refer to the specific description in the previous embodiment, which will not be repeated here.
[0085] Compared with related technologies, the method for associating ship SAR data and AIS data provided in this embodiment retains the technical effects of the aforementioned embodiments, and also verifies the target matching pair set before constructing the association relationship between SAR data and AIS data based on the target matching pair set, thereby improving the reliability of the target matching pair set and thus improving the reliability of the association results between ship SAR data and AIS data.
[0086] Based on the methods for correlating ship SAR data and AIS data, the corresponding methods include, for example: Figure 6 As shown in the embodiments of this application, a device for associating ship SAR data and AIS data is also provided. The device for associating ship SAR data and AIS data includes: The data acquisition module 601 acquires SAR data of all ships within a set area to form a SAR dataset, and acquires AIS data of all ships within the set area to form an AIS dataset. Topology building module 602 constructs a first ship topology based on a SAR dataset, which includes multiple SAR ship nodes, and constructs a second ship topology based on an AIS dataset, which includes multiple AIS ship nodes. The feature extraction module 603 is used to extract features from each SAR ship node to obtain SAR node feature vectors, and to extract features from each AIS ship node to obtain AIS node feature vectors. The cost matrix construction module 604 performs cross-operation on the SAR node feature vector and the AIS node feature vector to obtain the feature distance between each SAR ship node and each AIS ship node, and constructs a first-order cost matrix based on the feature distance. The matrix solving module 605 is used to solve the first-order cost matrix to obtain a set of first-order matching pairs. The set of first-order matching pairs is the set of matching pairs with the minimum sum of feature distances. The set of matching pairs includes multiple matching pairs, and each matching pair includes corresponding SAR data and AIS data. The data association module 606 is used to determine the target matching pair set based on the first-order matching pair set, and to construct the association relationship between SAR data and AIS data based on the target matching pair set.
[0087] The device for associating ship SAR data and AIS data provided in the above embodiments can realize the technical solutions described in the embodiments of the method for associating ship SAR data and AIS data. The specific implementation principles of each module or unit can be found in the corresponding content in the embodiments of the method for associating ship SAR data and AIS data, which will not be repeated here.
[0088] Please refer to Figure 7 This application also provides an electronic device 700. The electronic device 700 includes a processor 701, a memory 702, and a display 703. Figure 7 Only some components of the electronic device 700 are shown, but it should be understood that it is not required to implement all the components shown, and more or fewer components may be implemented instead.
[0089] In some embodiments, processor 701 may be a central processing unit (CPU), microprocessor, or other data processing chip, used to run program code stored in memory 702 or process data, such as the method for associating ship SAR data with AIS data in this application.
[0090] In some embodiments, processor 701 may be a single server or a group of servers. The server group may be centralized or distributed. In some embodiments, processor 701 may be local or remote. In some embodiments, processor 701 may be implemented on a cloud platform. In one embodiment, the cloud platform may include a private cloud, public cloud, hybrid cloud, community cloud, distributed cloud, intranet, multi-cloud, etc., or any combination thereof.
[0091] In some embodiments, memory 702 may be an internal storage unit of electronic device 700, such as a hard disk or memory of electronic device 700. In other embodiments, memory 702 may also be an external storage device of electronic device 700, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc. equipped on electronic device 700.
[0092] Furthermore, the memory 702 may include both internal storage units of the electronic device 700 and external storage devices. The memory 702 is used to store application software and various types of data installed on the electronic device 700.
[0093] In some embodiments, display 703 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. Display 703 is used to display information from electronic device 700 and to display a visual user interface. Components 701-703 of electronic device 700 communicate with each other via a system bus.
[0094] In one embodiment, when processor 701 executes the program for associating ship SAR data and AIS data in memory 702, the following steps can be implemented: SAR data of all ships within a specified area is acquired to form a SAR dataset, and AIS data of all ships within a specified area is acquired to form an AIS dataset. A first ship topology is constructed based on the SAR dataset, which includes multiple SAR ship nodes. A second ship topology is constructed based on the AIS dataset, which includes multiple AIS ship nodes. Feature extraction is performed on each SAR ship node to obtain the SAR node feature vector, and feature extraction is performed on each AIS ship node to obtain the AIS node feature vector. Cross-operation is performed on the SAR node feature vector and the AIS node feature vector to obtain the feature distance between each SAR ship node and each AIS ship node, and a first-order cost matrix is constructed based on the feature distance. Solving for the first-order cost matrix yields a set of first-order matching pairs. This set of first-order matching pairs is the set of matching pairs with the minimum sum of feature distances. The set of matching pairs includes multiple matching pairs, and each matching pair includes corresponding SAR data and AIS data. The target matching pair set is determined based on the first-order matching pair set, and the correlation between SAR data and AIS data is constructed based on the target matching pair set.
[0095] It should be understood that when the processor 701 executes the program for associating ship SAR data and AIS data in the memory 702, in addition to the functions mentioned above, it can also perform other functions, as detailed in the description of the corresponding method embodiments above.
[0096] Furthermore, this application does not specifically limit the type of electronic device 700 mentioned in the embodiments. Electronic device 700 can be a mobile phone, tablet computer, personal digital assistant (PDA), wearable device, laptop computer, or other portable electronic device. Exemplary embodiments of portable electronic devices include, but are not limited to, portable electronic devices running iOS, Android, Microsoft, or other operating systems. The aforementioned portable electronic device can also be other portable electronic devices, such as a laptop computer with a touch-sensitive surface (e.g., a touch panel). It should also be understood that in some other embodiments of this application, electronic device 700 may not be a portable electronic device, but rather a desktop computer with a touch-sensitive surface (e.g., a touch panel).
[0097] Accordingly, this application also provides a computer-readable storage medium for storing computer-readable programs or instructions. When the programs or instructions are executed by a processor, they can implement the steps or functions in the methods for associating ship SAR data and AIS data provided in the above-described method embodiments.
[0098] The table below shows the accuracy of the association results between ship SAR data and AIS data in different ship-dense scenarios (unit: %). As can be seen from the table, the association results of the ship SAR data and AIS data association method, device, electronic equipment and computer-readable storage medium provided in this application have high accuracy.
[0099]
[0100] in, SAR Matching Rate (MR) is the coverage rate of ship targets in a SAR image that are successfully matched to AIS targets, reflecting the algorithm's utilization of SAR ship targets. A higher SAR matching rate is better, as it means more SAR ship target data are matched and less SAR ship target data is wasted. The specific calculation formula is as follows: .
[0101] The AIS match rate is the coverage rate of AIS data that is successfully matched to SAR targets, reflecting the algorithm's utilization of AIS data. A higher AIS match rate is better, as it means more AIS data matches and less AIS data is wasted, which is crucial for resource allocation in maritime supervision. The specific calculation formula is as follows: .
[0102] The overall matching rate is the average of the SAR matching rate and the AIS matching rate. It avoids any single indicator being too high or too low, reflecting the algorithm's stability under multi-source data fusion. A higher overall matching rate is better, as it means more SAR and AIS ship target data are matched, resulting in better matching performance from the algorithm. The specific calculation formula is as follows: .
[0103] Precision is the probability of a correct match in the matching results, used to evaluate the accuracy of the algorithm's matching results. Higher precision is better, as it means more correctly matched SAR-AISAR node pairs. The specific calculation formula is as follows: .
[0104] Recall is the proportion of correct matches out of all true matches, used to evaluate an algorithm's ability to cover true targets. A higher recall is better, as it means fewer true targets are missed, which is crucial for safety-critical scenarios such as tracking illegal vessels. The specific formula for calculating recall is: .
[0105] The False Match Rate (FMR) is the proportion of incorrect matches out of all matches, reflecting the reliability of the matching results. A lower FMR is better, as it means the algorithm has a lower risk of misassociating data and avoids mistakenly tracking legitimate vessels. The specific formula for calculating FMR is as follows: .
[0106] The F1 score is the harmonic mean of precision and recall. A higher F1 score is better, as it indicates that the algorithm has achieved an optimal balance between precision and accuracy, resulting in the best overall performance. The specific calculation formula is as follows: .
[0107] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.), and the computer program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.
[0108] The foregoing provides a detailed description of the method, apparatus, electronic device, and storage medium for associating ship SAR data and AIS 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 core ideas of this application. At the same time, those skilled in the art will recognize that 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 method for correlating ship SAR data and AIS data, characterized in that, include: SAR data of all ships within a defined area are acquired to form a SAR dataset, and AIS data of all ships within the defined area are acquired to form an AIS dataset. A first ship topology is constructed based on the SAR dataset, the first ship topology including multiple SAR ship nodes; a second ship topology is constructed based on the AIS dataset, the second ship topology including multiple AIS ship nodes. Feature extraction is performed on each of the SAR ship nodes to obtain SAR node feature vectors, and feature extraction is performed on each of the AIS ship nodes to obtain AIS node feature vectors. The cross operation is performed on the SAR node feature vector and the AIS node feature vector to obtain the feature distance between each SAR ship node and each AIS ship node, and a first-order cost matrix is constructed based on the feature distance. Solving the first-order cost matrix yields a first-order matching pair set, which is the set of matching pairs with the smallest sum of feature distances. The matching pair set includes multiple matching pairs, and each matching pair includes corresponding SAR data and AIS data. Based on the first-order matching pair set, a target matching pair set is determined, and the association between the SAR data and the AIS data is constructed according to the target matching pair set.
2. The method for correlating ship SAR data and AIS data according to claim 1, characterized in that, Determining the target matching pair set based on the first-order matching pair set includes: The sum of feature distances of the first-order matching pair set is taken as the sum of target feature distances; A dynamic coefficient is set based on the sum of the distances to the target features; The second-order cost matrix is obtained by adjusting the first-order cost matrix according to the dynamic coefficients. Solve the second-order cost matrix to obtain the second-order matching pair set, and use the second-order matching pair set as the target matching pair set.
3. The method for correlating ship SAR data and AIS data according to claim 2, characterized in that, The step of setting dynamic coefficients based on the sum of the target feature distances includes: Based on formula The dynamic coefficients are calculated. in, 、 To set a constant, The dynamic coefficient is... It is the sum of the distances to the target features.
4. The method for correlating ship SAR data and AIS data according to claim 2, characterized in that, The step of adjusting the cost matrix according to the dynamic coefficients includes: Based on formula group ， Adjust the first-order cost matrix; in, For each element in the second-order cost matrix, For each element in the first-order cost matrix, The dynamic coefficient is... For the first The aforementioned SAR ship nodes For the first The aforementioned AIS ship nodes For the first topology The set of all adjacent nodes of the SAR ship node. For the second topology The set of all adjacent nodes of the AIS ship node.
5. The method for correlating ship SAR data and AIS data according to claim 4, characterized in that, The construction of the first ship topology based on the SAR dataset includes: For a target SAR ship node, the spatial distance between other SAR ship nodes and the target SAR ship node is calculated based on the SAR data. The SAR ship node corresponding to the target spatial distance is taken as the adjacent node of the target SAR ship node. The target SAR ship node can be any SAR ship node, and the target spatial distance is the spatial distance that meets the set requirements. The target SAR ship node is unidirectionally connected to the adjacent node to form the first ship topology.
6. The method for correlating ship SAR data and AIS data according to claim 1, characterized in that, The step of constructing the association between the SAR data and the AIS data based on the target matching pair set includes: For any target matching pair, determine whether the SAR data and the AIS data in the target matching pair are uniquely matched; If the SAR data and AIS data in the target matching pair are not uniquely matched, the target matching pair is removed from the target matching pair set.
7. The method for correlating ship SAR data and AIS data according to claim 1, characterized in that, The step of constructing the association between the SAR data and the AIS data based on the target matching pair set includes: For any of the target matching pairs, the actual distance between the ships corresponding to the SAR data and the ships corresponding to the AIS data is calculated based on the SAR data and the AIS data in the target matching pair; If the actual distance of the vessel does not meet the set conditions, the target matching pair is removed from the target matching pair set.
8. A device for correlating ship SAR data and AIS data, characterized in that, include: The data acquisition module is used to acquire SAR data of all ships within a set area to form a SAR dataset and to acquire AIS data of all ships within the set area to form an AIS dataset. A topology construction module is used to construct a first ship topology based on the SAR dataset, the first ship topology including multiple SAR ship nodes, and to construct a second ship topology based on the AIS dataset, the second ship topology including multiple AIS ship nodes. The feature extraction module is used to extract features from each of the SAR ship nodes to obtain SAR node feature vectors, and to extract features from each of the AIS ship nodes to obtain AIS node feature vectors. The cost matrix construction module is used to perform cross operations on the SAR node feature vector and the AIS node feature vector to obtain the feature distance between each SAR ship node and each AIS ship node, and construct a first-order cost matrix based on the feature distance. The matrix solving module is used to solve the first-order cost matrix to obtain a first-order matching pair set. The first-order matching pair set is the matching pair set with the smallest sum of feature distances. The matching pair set includes multiple matching pairs, and each matching pair includes corresponding SAR data and AIS data. The data association module is used to determine a target matching pair set based on the first-order matching pair set, and to construct the association relationship between the SAR data and the AIS data according to the target matching pair set.
9. An electronic device, characterized in that, Including memory and processor, among which, The memory is used to store programs; The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps in the method for associating ship SAR data and AIS data according to any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that, Used to store computer-readable programs or instructions, which, when executed by a processor, can implement the steps in the method for associating ship SAR data with AIS data as described in any one of claims 1 to 7.