False AIS signal detection method based on multi-source data

Abstract

The invention relates to the technical field of ship traffic management and marine monitoring, and discloses a multi-source data-based false AIS signal detection method, which comprises the following steps of: firstly, extracting a historical track to construct a coverage grid mapping table containing radar detection attributes; receiving real-time data and dividing the real-time data into an AIS to-be-detected group and a radar truth value group; non-radar blind area targets are screened according to the mapping table; retrieving an associated target in the radar truth value group to execute existence reverse verification; a dynamic time warping algorithm is adopted to align time sequences and calculate feature differences, and fine comparison verification is executed; constructing a neighborhood flow field model, and executing flow field consistency verification according to the motion vector deviation; and outputting a false AIS signal list in combination with multi-stage verification results. According to the method, through physical coverage modeling and environmental dynamics constraint, high-simulation-degree false signals in a radar blind area can be effectively identified, so that the false alarm rate is reduced.

Description

Technical Field

[0001] This invention relates to the field of ship traffic management and maritime surveillance technology, specifically a method for detecting false AIS signals based on multi-source data. Background Technology

[0002] Automatic Identification System (AIS), as a shipborne navigation and communication device, broadcasts dynamic information such as identity, location, speed, and heading to enable surrounding vessels and shore-based facilities to perceive traffic conditions, and is widely used in ensuring navigational safety and assisting traffic management. However, due to the open nature of the AIS protocol and the lack of a strict authentication mechanism, the system is vulnerable to spoofing attacks. Non-cooperative targets can generate non-existent false targets or tamper with real trajectories by forging or altering AIS signals, disrupting shipping order, and even being used to cover up illegal activities such as smuggling and illegal fishing.

[0003] Existing techniques for identifying fake AIS signals primarily rely on the feature analysis of the AIS data itself. For example, they analyze the kinematic rationality or signal reception range of ship dynamic data for identification. However, such methods struggle to counter highly realistic trajectory deception; attackers can simulate false trajectories that conform to motion patterns to evade detection. Another common approach is time slot conflict analysis based on the SOTDMA protocol. However, in sparsely populated sea areas with abundant time slot resources, attackers can easily select idle time slots to send fake packets without causing conflicts. In densely populated sea areas, normal time slot conflicts occur frequently, making it difficult to effectively distinguish between attack behavior and normal network congestion, resulting in a high false positive rate. Relying solely on a single AIS data source for verification is insufficient to accurately identify meticulously forged fake signals in the absence of external physical truth references. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method for detecting false AIS signals based on multi-source data, which solves the technical problems of existing technologies in distinguishing between radar blind spots and false signals in ship monitoring, as well as the difficulty in identifying false trajectories based on multi-dimensional features.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solution: a method for detecting false AIS signals based on multi-source data, comprising: establishing a radar coverage grid mapping table through mining historical data, which essentially constitutes a physical field-of-view map of the radar system, used to eliminate blind zone interference caused by objective factors such as terrain obstruction or distance limitations before real-time detection, and ensuring that only targets that should theoretically be detected are verified.

[0006] In the real-time detection process, this method establishes a verification logic based on radar observation data as the truth benchmark. For AIS targets within the non-blind zone, the system employs a three-level verification mechanism, from coarse to fine and from individual to environmental parameters.

[0007] The first level is existence reverse verification, which uses spatial proximity to quickly determine whether there are physical radar echoes around the AIS target in order to identify completely fictitious ghost ship signals.

[0008] The second level is spatiotemporal feature comparison and verification. A dynamic time warping algorithm is introduced to solve the problem of asynchronous sampling frequencies of heterogeneous sensors. Based on time alignment, the algorithm accurately calculates the subtle differences in the target's position, ship length, speed and heading to identify trajectory forgery or attribute tampering.

[0009] The third level is flow field consistency verification, which uses real radar targets in the neighborhood to construct a dynamic flow field model that reflects the current hydrological characteristics of the sea area. The motion vector of the target to be measured is compared with the environmental flow field to identify complex deceptive signals that, although the trajectory is smooth, violate the laws of ocean current movement.

[0010] Finally, the system integrates the verification results from the above three dimensions and outputs a list of high-confidence fake AIS signals.

[0011] This invention provides a method for detecting fake AIS signals based on multi-source data. It has the following beneficial effects:

[0012] 1. This invention quantifies the radar detection boundary by constructing a radar coverage grid mapping table that includes the minimum detection vessel length attribute. Targets located in radar blind zones or smaller than the radar's detection accuracy are eliminated before detection, avoiding the misclassification of real targets that cannot be detected due to physical limitations as false signals and reducing the system's false alarm rate.

[0013] 2. This invention employs a dynamic time warping algorithm to process heterogeneous data, resolving the issue of asynchronous sampling between AIS and radar. By performing fine-grained multi-dimensional feature comparison on the aligned trajectory sequences, and confirming the existence of echoes, it further identifies spurious trajectories where the position data exists but the motion characteristics do not match physical measurements, thus improving detection accuracy.

[0014] 3. This invention constructs a real-time flow field model based on neighborhood ground truth targets and introduces environmental dynamic constraints. By calculating the deviation between the motion vector of the target and the average vector of the local flow field, it identifies abnormal targets that conform to kinematic logic but violate hydrological laws; moreover, this method does not require pre-set hydrological data, can adapt to dynamic sea conditions, and enhances detection robustness. Attached Figure Description

[0015] Figure 1 This is an overall flowchart of the method according to several embodiments of the present invention;

[0016] Figure 2 This is a system functional module architecture diagram provided for one embodiment of the present invention. Detailed Implementation

[0017] The technical solutions in the embodiments of the present invention will be clearly and completely described below. 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.

[0018] See attached document Figure 2 This invention provides a method for detecting fake AIS signals based on multi-source data, comprising: a data acquisition module, a historical modeling module, a grid database, a real-time trajectory processing module, and a comprehensive detection module.

[0019] The data acquisition module is used to access and parse multi-source vessel data. This multi-source data includes Automatic Identification System (AIS) data and radar target data. The module performs format cleaning on the accessed data, extracting key fields including longitude, latitude, timestamp, vessel identification code, length, speed, and heading, and then divides the data into historical datasets and real-time data streams.

[0020] The historical modeling module is connected to the data acquisition module and is used to process radar signals in historical datasets. The historical modeling module extracts ship trajectory points with radar signals from historical data, performs GeoHash spatial encoding on the trajectory points, calculates the total amount of radar data and the minimum ship length attribute within each grid cell, and constructs a coverage model that reflects the radar detection capability.

[0021] The grid database is connected to the historical modeling module and is used to store statistical data on radar coverage grids. The grid database maintains a mapping table indexed by GeoHash encoding, which records the radar coverage status of each grid area and the lowest historically detected boat length value in that area.

[0022] The real-time trajectory processing module receives the real-time data stream output by the data acquisition module. It maintains a trajectory sequence for each ship target within a preset time window and, based on the data source attributes, divides the real-time data into two groups: a test group containing only AIS signals and a true value group containing radar signals.

[0023] The integrated detection module is connected to both the grid database and the real-time trajectory processing module to execute the logic for identifying false signals. The integrated detection module performs non-radar blind zone determination based on the grid database, spatiotemporal correlation reverse verification based on the ground truth data, trajectory similarity comparison, and neighborhood flow field consistency verification on the data to be tested, and outputs the final list of identified false AIS signals.

[0024] See attached document Figure 1 The method for detecting fake AIS signals includes the following steps:

[0025] S1. Extract ship trajectory data with radar signals from historical multi-source data, and construct a radar coverage grid mapping table containing radar detection capability attributes through GeoHash encoding.

[0026] S2. Receive real-time multi-source ship trajectory data, construct trajectory sequences according to ship identification codes, and divide the data into AIS test group and radar true value group according to data source attributes;

[0027] S3. Read the radar coverage grid mapping table, perform grid matching on the target trajectory in the AIS test group, and determine whether the target is in a non-radar blind zone based on the matching result.

[0028] S4. Perform spatiotemporal correlation between the AIS target group and the radar true value group targets that are determined to be in the non-radar blind zone, and perform existence reverse verification based on whether there is radar echo within the surrounding preset range.

[0029] S5. If there are radar true value group targets within the preset range, use the dynamic time warping algorithm to align the time series and calculate the trajectory feature differences to perform comparison and verification.

[0030] S6. Select the radar true value group of targets within the neighborhood to construct the flow field model, and perform flow field consistency verification based on the deviation between the motion vector of the AIS target group and the average vector of the neighborhood flow field.

[0031] S7. Based on the judgment results of the above verification steps, output a list of identified fake AIS signals.

[0032] Step S1: Establish radar coverage feature grid.

[0033] This step quantifies the detection capability boundaries of radar in different geographical areas by analyzing the statistical patterns of historical data, providing a true reference for subsequent reverse verification.

[0034] The system accesses a historical multi-source ship trajectory database stored on a server. This historical data covers a massive number of trajectory points within a preset time span. The system performs filtering operations to select trajectory data with clear radar signal markers. Data with clear radar signal markers includes single-source radar tracking data and fused trajectory data that has undergone multi-source fusion and includes radar components. For each selected trajectory record, the system extracts its spatial location and physical attributes to construct a feature vector set. eigenvectors Indicates the first The data for each trajectory point is defined as follows:

[0035] ;

[0036] in, Indicates longitude. Indicates latitude, Indicates the sampling time. This indicates the ship's length. For single-source radar data, if direct ship length information is lacking, the system retrieves the corresponding ship's design length based on the MMSI or batch number of the radar-tracked target by associating it with the historical AIS static database; if no association is possible, the data entry is discarded.

[0037] The system uses the GeoHash algorithm to process the longitude in the above feature vectors. and latitude Spatial discretization encoding is performed. The GeoHash algorithm maps two-dimensional latitude and longitude coordinates to one-dimensional strings using a binary search method. The length of the string determines the precision of the grid. In this embodiment, the encoding precision is set to Level 7 (i.e., the string length is 7 bits). At this precision, each GeoHash grid represents a rectangular area of ​​approximately 153 meters × 153 meters on the Earth's surface.

[0038] The system uses GeoHash encoded values As the aggregation key, feature vectors falling within the same grid are grouped and statistically analyzed. For each grid... Calculate the following two core statistical indicators:

[0039] First, the density of radar locations This represents the total number of radar echo points falling within this grid in historical data, used to characterize the coverage frequency of radar signals in this area;

[0040] Second, the minimum length of the probe ship. This represents the smallest ship size that has historically been detectable by radar within the grid.

[0041] Minimum Detection Ship Length The calculation formula is:

[0042] ;

[0043] in, This indicates all items falling into the grid. The set of captains corresponding to the trajectory points. This is a function for finding the minimum value. Furthermore, in another embodiment, to eliminate the influence of individual abnormal noise points, the minimum detection length can also be determined using the percentile method in statistics. For example, the value at a predetermined percentile (such as the 5th or 10th percentile) in the set of lengths can be calculated as the minimum detection length for that grid. Introducing the minimum detection length attribute can establish the lower limit of the radar's detection capability in this area, preventing the misclassification of real small vessels that the radar cannot detect due to physical limitations as false AIS signals during subsequent real-time detection.

[0044] The system stores the calculated statistical indicators into a grid database and establishes a key-value pair mapping relationship. The mapping relationship is represented as follows:

[0045] ;

[0046] Finally, a mapping relationship is generated. The radar coverage grid mapping table is persistently stored and can be loaded into memory for quick querying by the real-time detection module.

[0047] Step S2: Real-time trajectory construction and grouping.

[0048] The real-time trajectory processing module continuously receives multi-source ship data streams. The system adopts a hash table-based in-memory storage structure, using the ship's unique identifier as the key to maintain the real-time trajectory sequence of each ship. System-defined time window A sliding window management system is implemented for real-time trajectory sequences. Consisting of a series of ordered data points Composition, represented as:

[0049] ;

[0050] The system iterates through all trajectory sequences within the current time window and divides them into two groups based on the data source identifier attribute: the AIS test group. and radar truth set The partitioning logic is as follows: if a trajectory sequence contains only data from AIS base stations, then it is classified as... If a trajectory sequence contains data from a radar station, or is a fused trajectory generated by fusing AIS data and radar data (i.e., the target has been captured by a radar sensor), then it is classified as... .

[0051] Step S3: Mesh matching and blind spot determination.

[0052] The system from Select one target vessel to be inspected. Obtain its trajectory sequence And the captain attribute in static information .for Each trajectory point in The system calculates its corresponding GeoHash code. And utilize the mapping relationship of this encoding in memory loading. The query is performed on the radar coverage grid mapping table. The query operation aims to determine whether a point meets the conditions that it should be detected by radar. Specific determination conditions are as follows: Defined as:

[0053] ;

[0054] in, This represents the set of all grid key-value pairs that contain records in the mapping table. This indicates the minimum historical probe length recorded by this grid.

[0055] After completing the matching status marking of all points in the trajectory sequence, the system calculates the coverage of the trajectory sequence within the radar's effective detection area. :

[0056] ;

[0057] System set coverage threshold .like Then the AIS target is determined. Currently sailing within a non-radar blind zone, theoretically, a radar echo signal should be generated; conversely, if... If the target is not detected, it is considered to be in a radar blind zone or at the edge of radar detection capability, and the system will not make any further false signal determinations.

[0058] Step S4: Spatiotemporal correlation and reverse verification.

[0059] For AIS targets determined to be outside the radar blind zone The system performs a reverse existence verification. The system uses... Current position Set the search radius around the center. (e.g., 200 meters), in the radar truth set The search query determines whether a target satisfying the spatial distance constraint exists. The search criteria are expressed as follows:

[0060] ;

[0061] in, The location of the target in the radar truth set. This is a function for calculating the geographic Euclidean distance between two points.

[0062] Based on the search results, the system performs the following logical judgment: if within the search radius... No results were found belonging to [the relevant category]. If the target, i.e. the search results, is an empty set, then the AIS target is determined to be empty. This is a false signal lacking physical echo support. The basis for this determination is that for a target that meets radar detection conditions (i.e., is in a non-blind zone and its size meets the detection threshold), there should be corresponding radar target data at its location coordinates. If the search result is not empty, meaning there is at least one radar target nearby, the system will pass the AIS target and the searched radar target candidate set to the next-level comparison module for subsequent trajectory similarity verification.

[0063] Step S5: Trajectory similarity comparison.

[0064] Once the system has detected a set of candidate radar targets around an AIS target, it will process each pair of AIS target trajectory sequences. and radar target trajectory sequence The system performs time synchronization. Because the sampling frequencies of the AIS and radar systems are not synchronized, the system uses a dynamic time warping algorithm to establish the optimal matching path between the two time series. The system constructs a cost matrix and uses dynamic programming to find a path that minimizes the total time difference cost, outputting a set of time-aligned trajectory point pairs. .

[0065] Based on the aligned point pairs, the system calculates the feature differences across three dimensions: spatial location, physical attributes, and motion state. For each pair of points, the following difference indices are calculated:

[0066] Location distance difference : Calculate the geographic Euclidean distance between two latitude and longitude coordinate points;

[0067] The captain's attributes are poor. : Calculate the absolute value of the difference between the AIS reported length and the radar estimated length. ;

[0068] Speed ​​difference Calculate the difference between the AIS reported speed and the radar observed speed. ;

[0069] Heading differences To calculate the difference in headings, taking into account the periodicity of angles, the formula is:

[0070] ;

[0071] The system calculates the arithmetic mean of the four indicators mentioned above over the entire trajectory sequence to obtain the average distance difference. Average ship length difference Average speed difference and average heading difference The system sets a set of similarity judgment thresholds. If any one or more indicators exceed the threshold, the radar target is determined not to be the physical entity corresponding to the AIS target. If none of the candidate radar targets meet the similarity condition after traversing all candidates, the system identifies the AIS target as a false signal.

[0072] Step S6: Neighborhood flow field consistency verification.

[0073] To further identify deceptive signals that conform to trajectory characteristics but violate environmental physical laws, this step constructs a real-time flow field model of a local sea area to verify whether the motion state of the AIS target under test conforms to the physical and dynamic laws of the current environment. Specifically, this includes three steps: constructing a flow field reference set, modeling flow field vectors, and determining the deviation.

[0074] For constructing the flow field reference set, the system uses the GeoHash grid where the AIS target under test is located at the current moment. Based on this, select the adjacent grid cells. Each grid serves as the flow field sampling region. Within the flow field sampling region, the system searches for all data belonging to the radar truth set. The target, along with the AIS targets that have been verified and marked as real in the previous steps, constitute the flow field reference dataset. .like If the number of targets is less than the preset statistical lower limit threshold, the system skips this step of verification.

[0075] For modeling the flow field vectors, the system uses a flow field reference dataset. The motion vectors of all targets are statistically aggregated to calculate the average flow field vector for that local sea area. Given the periodicity of the heading angle data, the system employs vector decomposition and synthesis for calculation. For the flow field reference dataset... Each reference target in The system obtains its ground speed. and ground heading Based on trigonometric relationships, the target's motion vector is decomposed into an eastward component. and due north component :

[0076] ;

[0077] ;

[0078] Based on the above decomposition results, the system calculates the average eastward component of all reference targets within the region. and the average north component :

[0079] ;

[0080] ;

[0081] in, This represents the total number of targets in the reference dataset. Using the calculated average components, the system synthesizes the local flow field average vector. The magnitude of this vector is the flow velocity. and direction, i.e., flow direction The results are obtained using the following formulas:

[0082] ;

[0083] ;

[0084] In calculating the flow direction At that time, the system determines the quadrant of the angle based on the sign of the average component and maps the result to the range of 0 to 360 degrees. This local flow field average vector characterizes the background motion trend in the current sea area under the combined influence of environmental factors such as ocean currents and wind.

[0085] To determine the deviation, the system acquires the motion vector of the AIS target under test. Including speed and heading The system establishes a consistency deviation index between the target object and the local flow field model. The definition is as follows:

[0086] ;

[0087] In the above formula, and The preset weighting coefficients, It is a non-zero, small constant. The system sets a flow field consistency threshold. .like Greater than If the motion state of the AIS target being tested deviates from the physical flow field law of the current sea area (for example, maintaining an ideal straight-line navigation unaffected in a strong rip current area), the system will identify it as a false AIS signal.

[0088] Step S7: Output the detection results.

[0089] The system executes comprehensive decision logic, summarizes the output states of the above verification steps, and generates the final detection result. The comprehensive decision logic follows a logical OR principle: if the AIS target under test meets any of the following conditions, the system determines that the target is a false signal and generates a corresponding false flag:

[0090] Condition 1: Reverse existence check failed;

[0091] Condition 2: Trajectory feature comparison failed;

[0092] Condition 3: The flow field consistency check failed.

[0093] The system generates a list of all false AIS signals identified as false targets. This list records the vessel identification code, the timestamp of the determination, and the specific status indicator used as the basis for the determination. The output interface sends the list of false AIS signals to the downstream vessel traffic management system or human-machine interface terminal. On the electronic chart display interface of the human-machine interface terminal, the system visually marks the targets in the list, highlighting them with a different color or line type than normal targets, and displays accompanying text prompts explaining the reasons for the determination, to assist maritime regulatory personnel in subsequent manual verification or enforcement actions.

[0094] This invention provides an electronic device that can act as an execution entity to implement the false AIS signal detection method based on reverse verification of radar signals and consistency of neighborhood flow fields in multi-source data as described in the above embodiments. The electronic device includes a processor, a memory, a communication interface, and an input / output interface connected via a system bus. The processor provides computational and control capabilities to support the operation of the entire system. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and a computer program. When the computer program is executed by the processor, it implements the aforementioned false AIS signal detection method. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The communication interface is used for network communication, such as accessing the Automatic Identification System (AIS) data link and radar signal data network to receive real-time multi-source trajectory data streams.

[0095] A processor, as the core computing unit of an electronic device, should be understood as a hardware circuit with data processing capabilities. In specific implementations, a processor can be a general-purpose central processing unit, microcontroller, digital signal processor, application-specific integrated circuit (ASIC), field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A processor can be a single-core processor or a multi-core processor. When executing program instructions stored in memory, a processor can perform various logical operations, including historical radar coverage grid modeling, real-time trajectory sequence construction and grouping, non-radar blind zone determination, spatiotemporal correlation reverse verification, trajectory similarity comparison, and neighborhood flow field consistency verification.

[0096] The memory is used to store data and instructions. In specific implementations, the memory may include high-speed random access memory or non-volatile memory, such as at least one disk storage device, flash memory device, general-purpose flash memory, etc. The computer program stored in the memory includes program instructions, which, when executed by the processor, cause the electronic device to perform the steps of the spoofed AIS signal detection method described in any of the above embodiments.

[0097] The system bus serves as a channel connecting various hardware components, used to transmit information between the processor, memory, and communication interfaces.

[0098] This invention also provides a computer-readable storage medium. This computer-readable storage medium may be non-transitory and stores a computer program thereon. When the computer program is executed by a processor, it implements the various steps of the false AIS signal detection method based on reverse verification of radar signals and consistency of neighborhood flow fields provided in the above embodiments of this invention. The computer-readable storage medium includes, but is not limited to, magnetic storage media, optical storage media, magneto-optical storage media, and semiconductor memories. Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods.

Claims

1. A method for detecting fake AIS signals based on multi-source data, characterized in that, include: Ship trajectory data with radar signals were extracted from historical multi-source data, and a radar coverage grid mapping table containing radar detection capability attributes was constructed using GeoHash encoding. Receive real-time multi-source ship trajectory data, construct a trajectory sequence based on the ship identification code, and divide the trajectory sequence into an AIS test group and a radar true value group based on the data source attributes of the data points in the trajectory sequence; Based on the radar coverage grid mapping table, the trajectories in the AIS target group are matched to filter out the ship targets to be detected that are not in the radar blind zone; Search the radar truth set for associated radar targets around the target ship, and determine the existence of radar echoes of the target ship based on the search results, as a reverse verification of existence. In the case of the retrieved associated radar targets, a dynamic time warping algorithm is used to perform spatiotemporal alignment and feature difference calculation on the ship target to be detected and each of the retrieved associated radar targets. Based on the feature difference calculation results, the physical identity between the ship target to be detected and the associated radar targets is determined as a comparison verification. A flow field model is constructed by selecting radar true value group targets within the neighborhood of the target ship to be detected, and flow field consistency verification is performed based on motion vector deviation. Based on the results of the existence reverse verification, the comparison verification, and the flow field consistency verification, a list of fake AIS signals is output.

2. The method for detecting fake AIS signals based on multi-source data according to claim 1, characterized in that, The construction of a radar coverage grid mapping table containing radar detection capability attributes using GeoHash encoding includes: The longitude and latitude of the historical trajectory data with radar signal markers extracted from the historical multi-source data are spatially discretized and encoded using GeoHash. Using GeoHash encoded values ​​as aggregation keys, the radar point density and minimum detection ship length within each grid are statistically analyzed. The minimum probe length is obtained by selecting one of the following methods: calculating the minimum value of the set of lengths corresponding to all trajectory points falling within the grid and calculating the value of the set of lengths at a predetermined percentile. The GeoHash encoded value is used as the key, and the radar point density and the minimum detection ship length are used as the values, and stored in the radar coverage grid mapping table.

3. The method for detecting fake AIS signals based on multi-source data according to claim 1, characterized in that, Based on the data source attributes of the data points in the trajectory sequence, the trajectory sequence is divided into an AIS test group and a radar truth group, including: For each ship, maintain the trajectory sequence within the predetermined time window; If a trajectory sequence contains only data from an AIS base station, the trajectory sequence is assigned to the AIS test group. In the case where a trajectory sequence contains data from a radar station, and in the case where a trajectory sequence contains fused trajectory data generated by fusing AIS data and radar data, the trajectory sequence is assigned to the radar truth set.

4. The method for detecting fake AIS signals based on multi-source data according to claim 2, characterized in that, The process of matching trajectories in the AIS target group based on the radar coverage grid mapping table to filter out target ships in non-radar blind zones includes: Obtain the trajectory sequence of the target to be screened in the AIS test group and the captain key field contained in the trajectory sequence; Calculate the GeoHash code for each trajectory point in the trajectory sequence and query it in the radar coverage grid mapping table; If the GeoHash code corresponding to the trajectory point exists in the radar coverage grid mapping table, and the value of the ship length key field exceeds the minimum detection ship length recorded in the grid, then the trajectory point is determined to meet the radar detection conditions. The coverage rate of trajectory points that meet the radar detection conditions in the entire trajectory sequence is calculated. If the coverage rate exceeds a predetermined threshold, the target to be screened is determined to belong to the ship target to be detected in the non-radar blind zone.

5. The method for detecting fake AIS signals based on multi-source data according to claim 4, characterized in that, The existence reverse verification includes: Using the position corresponding to the latest timestamp in the trajectory sequence of the ship to be detected as the center, the target in the radar truth set is retrieved within a predetermined search radius; If the search results are empty, it is determined that the target vessel to be detected belongs to a false AIS signal; If the search results are not empty, the set of all retrieved targets is taken as the candidate set of associated radar targets and proceeds to the comparison and verification step.

6. The method for detecting fake AIS signals based on multi-source data according to claim 5, characterized in that, The step of using a dynamic time warping algorithm to perform spatiotemporal alignment between the target ship to be detected and each of the retrieved associated radar targets includes: For the trajectory sequence of the ship target to be detected and the trajectory sequence of any target in the associated radar target candidate set, a cost matrix is ​​constructed, where the elements of the cost matrix represent the absolute value of the timestamp difference between the data points in the two sequences. The dynamic programming method is used to find the path that passes through the cost matrix and minimizes the total time difference cost, and outputs a set of time-aligned trajectory point pairs.

7. The method for detecting fake AIS signals based on multi-source data according to claim 6, characterized in that, The determination of the physical identity between the target ship and the associated radar target based on the feature difference calculation results, as a comparison verification, includes: Based on the time-aligned trajectory point pair sequence, the average position distance difference, the average ship length attribute difference, the average speed difference, and the average heading difference are calculated respectively. A predetermined similarity threshold is determined. If any one of the average difference in position distance, the average difference in ship length attribute, the average difference in speed, and the average difference in heading exceeds the corresponding predetermined similarity threshold, it is determined that any target is not the physical entity corresponding to the ship target to be detected. If none of the targets in the associated radar target candidate set meet the similarity condition, the ship target to be detected is identified as a fake AIS signal.

8. The method for detecting fake AIS signals based on multi-source data according to claim 1, characterized in that, The step of selecting radar true value group targets within the neighborhood of the target ship to be detected to construct a flow field model includes: The grid corresponding to the latest timestamp in the target trajectory sequence of the ship to be detected is used as the reference, and the grids adjacent to the reference are selected as the flow field sampling area; Within the flow field sampling area, all targets belonging to the radar true value group and AIS targets that have been verified as real in previous detection cycles are retrieved to form a flow field reference dataset; If the number of targets in the flow field reference dataset does not reach the predetermined statistical lower limit threshold, the flow field consistency check is skipped.

9. A method for detecting fake AIS signals based on multi-source data according to claim 8, characterized in that, The calculation of the neighborhood flow field average vector in the constructed flow field model includes: Obtain the ground speed and ground heading of each reference target in the flow field reference dataset; The motion vector of each reference target is decomposed into an eastward component and a northward component using trigonometric functions. Calculate the arithmetic mean of the eastward and northward components of all reference targets within the region, respectively. Based on the arithmetic mean of the eastward component and the arithmetic mean of the northward component, a local flow field average vector is synthesized, which includes flow velocity and flow direction.

10. A method for detecting fake AIS signals based on multi-source data according to claim 9, characterized in that, The flow field consistency verification based on motion vector deviation includes: The speed and heading of the target vessel are obtained to form a motion vector; The consistency deviation index is calculated, which is obtained by weighted summation of speed difference and heading difference terms; The speed difference term refers to the ratio of the absolute value of the difference between the target speed of the vessel under test and the average vector velocity of the local flow field to the velocity itself. The heading difference term is calculated based on the cosine similarity between the target heading of the vessel to be detected and the average vector direction of the local flow field; If the consistency deviation index exceeds the predetermined flow field consistency threshold, the ship target to be detected is determined to be a false AIS signal.

Images