Radar plot and AIS track adaptive parallel matching method and system

Abstract

The invention discloses a radar plot and AIS track adaptive parallel matching method and system, and belongs to the technical field of radar and navigation information. In order to reduce the algorithm complexity and the dependence of AIS data, the method comprises the steps that database arrival data are identified, a system error extraction module votes radar plots after projection in a parameter space based on an error distribution function to obtain a plurality of counters, and the counters are filtered through a filter; finally, results of all filters are accumulated, then peak detection is carried out, and the system error of the radar plot is extracted. And the radar plot index module performs speed relevance check on the retrieved radar track points to retrieve the radar track points of which the speed parameters meet the requirements. And the error function optimization module performs sum optimization on the error function according to a PARZEN window based on the obtained radar track points of which the speed parameters meet the requirements. And updating the number of the subordinate ship in the database AIS table by using the radar track point with the speed parameter meeting the requirement.

Description

Technical Field

[0001] The present invention belongs to the technical fields of radar and marine navigation information technology, and particularly relates to a method and system for adaptively and parallelly matching radar traces and AIS tracks. Background Art

[0002] With the increasing busyness and complexity of marine traffic, the density of ship traffic in coastal, port and inland water areas is increasing day by day. Relying solely on the AIS vessel management system has problems under high load.

[0003] Traditional ship traffic management systems mainly rely on two subsystems: AIS and radar systems.

[0004] AIS data has the characteristics of high precision and wide coverage; however, the arrival of AIS data is unpredictable and random. Especially in high-density traffic areas, especially the AIS data near the shore usually uses the Self-Organized Time Division Multiple Access (SODTMA) technology. When the slot resources are insufficient or there is a shortage of slot resources, it may lead to the loss of ship data. In addition, the revisit interval of AIS data may be very large.

[0005] Correspondingly, radar is actively detected, and the time interval of the radar trace data obtained therefrom is fixed and controllable; however, the radar system is significantly affected by the environment and has a relatively large position error compared to relying on AIS data. Usually, it is used in combination with AIS data to provide redundant detection capabilities and supplement in the case of missing or incorrect AIS data.

[0006] Traditional correlation methods, such as track-to-track methods, rely on first constructing a track for the radar trace data and then matching it with AIS data. This solution can effectively avoid the situation of reporting AIS data and reporting false AIS data. A large number of mature solutions have also been given in early literature, which can achieve the optimal fusion mode in any communication mode (such as periodic communication mode, event-triggered communication mode, full rate communication) and have a mature performance boundary estimation for involved fusion modification. However, this type of method has the problem of repeated utilization of heading and speed parameters during the process of track construction and matching. In addition, all such algorithms rely on time correlation, are difficult to parallelize, and due to the repeated matching of some information, naturally lead to a slow operation time of such algorithms.

[0007] Another correlation method relies on algorithms for measurement level fusion. Such algorithms rely on Joint Probabilistic Data Association (JPDA), can effectively perform data association in high noise, can handle complex scenarios when targets cross or approach, can effectively utilize multi-source information, and complete matching in complex scenarios. However, this type of method is very sensitive to missed detections and false alarms, and as the number increases, the hypothesis space will expand extremely rapidly, making it inapplicable to areas with a high density of ships such as coastal areas and ports. In addition, although the JPDA algorithm can be parallelized, the premise of its parallelization is that the movements of targets are independent of each other, which is also inapplicable to the matching fusion in port areas.

[0008] In summary, the two main matching methods in the prior art have paid a computational cost to achieve optimal matching, are not suitable for the complex environment in coastal areas, and both rely too much on AIS data. When AIS data is missing, a series of incorrect matches are likely to occur. Summary of the Invention

[0009] The problem to be solved by the present invention is to reduce the algorithm complexity and the dependence on AIS data, and propose a method and system for adaptive parallel matching of radar echoes and AIS tracks.

[0010] To achieve the above object, the present invention is realized through the following technical solutions:

[0011] A method for adaptive parallel matching of radar echoes and AIS tracks includes the following steps:

[0012] S1. Identify the data arriving at the database. The data identified as AIS data is saved in the AIS table of the database and undergoes data preprocessing, and the data identified as radar data is saved in the radar table of the database. When the saved radar data exceeds a certain batch, proceed to the next step;

[0013] S2. The system error extraction module interpolates the AIS data obtained in step S1 to align its time with the time of the radar data, then projects the radar echoes into the parameter space through the differential transformation method based on the longitude and latitude of the AIS data, then votes on the projected radar echoes in the parameter space based on the error distribution function to obtain multiple counters, filters the obtained multiple counters through a filter, and finally accumulates the results of all filters and then performs peak detection to extract the system error of the radar echoes.

[0014] S3. The radar track point indexing module retrieves radar track points based on the system error obtained in step S2, performs a velocity correlation check on the retrieved radar track points, proceeds to the next step for the radar track points with velocity parameters meeting the requirements, and retrieves the radar track points with non-compliant velocity parameters again until the specified number of times is reached or radar track points with velocity parameters meeting the requirements are retrieved;

[0015] S4. The error function optimization module optimizes the error function based on the radar track points with velocity parameters meeting the requirements obtained in step S3 according to the PARZEN window;

[0016] S5. Update the affiliated vessel number in the database AIS table based on the radar track points with velocity parameters meeting the requirements obtained in step S3.

[0017] Furthermore, the specific implementation method of step S1 includes the following steps:

[0018] S1.1. Set the data extracted from the database AIS table to include the Maritime Mobile Service Identity (MMSI), time, longitude, latitude, speed, navigation, and affiliated vessel number, and construct a uniqueness constraint for the composite key of MMSI and time in the AIS data;

[0019] S1.2. Set the data extracted from the database radar table to include the primary key, time, longitude, latitude, and radial velocity;

[0020] S1.3. When it is detected that there are more than three batches of existing radar data, proceed to the next step.

[0021] Furthermore, the specific implementation method of step S2 includes the following steps:

[0022] S2.1. Interpolate the AIS data obtained in step S1 to align its time with the time of the radar data, and use the second-order spline interpolation method to interpolate the AIS data to obtain the position of the AIS data at the radar scan time;

[0023] S2.2. Project the radar track points into the parameter space through the differential transformation method according to the longitude and latitude of the AIS data. The expression is:

[0024]

[0025] where P' radar,t is the position of the radar track point after differential transformation at time t, P ais,t is the position of the AIS data at time t, P radar,t is the position of the radar track point at time t, T radar is the time of radar scan, T ais,start is the time of the first point of the AIS data currently used for matching.ais,end is the time at which the last point of the AIS data used for performing the matching is located; it is set that the projected radar traces are represented by (x radar,t , y radar,t ).

[0026] S2.3. For the projected radar traces obtained in step S2.2, vote based on the error distribution function according to the Bayesian estimation framework, as shown in the following formula:

[0027] f(x - x radar,t , y - y radar,t )

[0028] where f(x, y) is the error distribution function, and x and y are the variables of the error distribution function;

[0029] The same vote is performed for each projected radar trace, and the result is expressed as:

[0030]

[0031] where D t is the set of projection points of all radar traces in the parameter space at time t, is the counter at time t, used to represent the voting result at time t;

[0032] S2.4. For the multiple counters obtained in step S2.3 filter them using the following filter:

[0033]

[0034] where H n,t represents the filtering result at time t, n represents the filter number, and N * represents the set of positive integers;

[0035] Based on the filtering result using a threshold of 0.05·n, accumulate the results of all filters, and the accumulated result is expressed as:

[0036]

[0037] For perform peak detection, set the deviation of the point with the highest value compared to the origin as the systematic error, and the detected peak is (x peak , y peak ).

[0038] Furthermore, the specific implementation method of step S3 includes the following steps:

[0039] S3.1. Radar Point Trajectory Indexing: The system error obtained in step S2 is transformed as follows:

[0040] P' ais,t = P ais,t + (x peak , y peak )

[0041] where P' ais,t is the reference point for indexing radar point trajectories at time t;

[0042] Using the nearest neighbor method, the closest point is selected for each radar point trajectory to obtain the radar track point P radar,t,0 , and the expression is:

[0043]

[0044] where argmin represents finding the point P t in the radar point trajectory set D radar,t that minimizes ||P ais,t - P' radar,t ||2. || ||2 represents taking the two-norm of the vector therein. Specifically, for (x1, x2,... x n ), this symbol means performing the operation using the following formula

[0045] S3.2. Velocity Correlation Check: For the radar track point P radar,t,0 output in step S3.1, a radial velocity correlation check is performed. If the radial velocity is within three times the radar's velocity resolution range, it is considered that the selection of this radar track point is correct, and the matching radar track point is output, and step S4 is executed; otherwise, step S3.3;

[0046] S3.3. For the radar track points that do not meet the requirements, this point is removed. For the remaining point set D t \{P radar,t,0}, find the new closest point P radar,t,1 as follows:

[0047]

[0048] Then, perform the velocity correlation check again. If it still does not meet the requirements, continue to search further in the set P radar,t ∈D t {P radar,t,0 , P radar,t,1} until the number of repetitions reaches K times, then stop searching, output all the radar track points obtained by indexing, and execute step S4 to obtain the optimal radar track point obtained by indexing as Pradar,t,max .

[0049] Further, the specific implementation method of step S4 includes the following steps:

[0050] S4.1. Optimize the error function of the corresponding radar block based on the radar track points indexed in step S3. The expression is as follows:

[0051]

[0052] where D k are all the indexed radar track points located in block i, with a total of k + 1. f i,old is the old error distribution function in block i, and f i,new is the updated error distribution function in block i. (lon P , lat P ) is the offset of the indexed radar track point P from the system error:

[0053] (lon P , lat P ) = P radar,t,max - P ais,t - (x peak , y peak );

[0054] S4.2. Re-store the updated error distribution function in step S4.1 into the hard disk in the form of a constructed data table.

[0055] A radar point track and AIS track adaptive parallel matching system is implemented based on the described radar point track and AIS track adaptive parallel matching method, and includes a database, a system error extraction module, a radar point track indexing module, and an error function optimization module;

[0056] The database is connected to the system error extraction module. The system error extraction module is respectively connected to the radar point track indexing module and the error function optimization module. The radar point track indexing module is also connected to the database.

[0057] Advantages of the present invention:

[0058] For the radar point track and AIS track adaptive parallel matching method of the present invention, by projecting the radar point tracks to be matched into the parameter space, the temporal sequence relationship of the radar point tracks is eliminated, and the traditional track building - matching problem is transformed into a parameter scanning problem. Therefore, this method has the ability to execute in parallel and does not depend on the temporal sequence, thus greatly improving the processing efficiency.

[0059] An adaptive parallel matching method for radar point traces and AIS tracks according to the present invention can continuously optimize the error distribution during the operation of the system and achieve more accurate and robust radar data processing by iteratively using the error function construction and peak index technology. At the same time, based on the adaptive error distribution function, this method avoids the computational cost and huge assumption space requirements for optimizing the matching results.

[0060] An adaptive parallel matching method for radar point traces and AIS tracks according to the present invention combines parameter space mapping and peak detection technology to achieve parallel and efficient extraction of radar system errors, so as to accurately index the traces matching the true tracks from the radar point traces. In addition, during the operation, the present invention can dynamically optimize the error distribution according to the indexed tracks to achieve adaptive matching optimization, significantly improving the robustness and real-time performance of radar data processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 is a flowchart of an adaptive parallel matching method for radar point traces and AIS tracks according to the present invention;

[0062] Figure 2 is a structural block diagram of an adaptive parallel matching method for radar point traces and AIS tracks according to the present invention;

[0063] Figure 3 is a schematic diagram of projecting radar point traces into the parameter space according to the present invention;

[0064] Figure 4 is a schematic diagram of the voting result in the parameter space according to the error distribution of the present invention;

[0065] Figure 5 is a schematic diagram of the optimized error distribution function using the PARZEN window in the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0066] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention, that is, the specific embodiments described are only a part of the embodiments of the present invention, rather than all of the specific embodiments. The components of the specific embodiments of the present invention usually described and shown in the drawings here can be arranged and designed in various different configurations, and the present invention can also have other embodiments.

[0067] Accordingly, the following detailed description of the specific embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely represents selected specific embodiments of the present invention. All other specific embodiments obtained by those skilled in the art based on the specific embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0068] To further understand the content, features and effects of the present invention, the following specific embodiments are exemplified and described in conjunction with the appended Figure 1 - appended Figure 5 as follows in detail:

[0069] Example 1:

[0070] A method for adaptively parallel matching of radar tracks and AIS tracks includes the following steps:

[0071] S1. Identify the data arriving at the database. Save the data identified as AIS data into the AIS table of the database and perform data preprocessing. Save the data identified as radar data into the radar table of the database. When the saved radar data exceeds a certain number of batches, proceed to the next step;

[0072] Further, the specific implementation method of step S1 includes the following steps:

[0073] S1.1. Set the data extracted from the AIS table of the database to include the Maritime Mobile Service Identity (MMSI), time, longitude, latitude, speed, navigation, and affiliated vessel number. Construct a uniqueness constraint for the composite key of MMSI and time in the AIS data;

[0074] The format of the database radar table is shown in Table 1:

[0075] Table 1 Database Radar Table

[0076]

[0077] The affiliated vessel number is allowed to be null, but for AIS that is null for a long time, it should be marked as -1;

[0078] S1.2. Set the data extracted from the database radar table to include the primary key, time, longitude, latitude, and radial velocity;

[0079] The format of the database radar table is shown in Table 2:

[0080] Table 2 Database Radar Table

[0081]

[0082] S1.3. When it is detected that the existing radar data is greater than three batches, proceed to the next step;

[0083] Furthermore, before the correct output result is obtained, this method needs to run a certain number of rounds in advance with the latest radar data and corresponding AIS data. It is recommended that the specific number of rounds be greater than twenty to ensure that the error distribution function is fully updated.

[0084] S2. The system error extraction module interpolates the AIS data obtained in step S1 to align its time with that of the radar data, then projects the radar traces into the parameter space according to the longitude and latitude of the AIS data through the differential transformation method, and then votes on the projected radar traces in the parameter space based on the error distribution function to obtain multiple counters. The obtained multiple counters are filtered through a filter, and finally the results of all filters are accumulated and then peak detection is performed to extract the system error of the radar traces;

[0085] Furthermore, the specific implementation method of step S2 includes the following steps:

[0086] S2.1. Interpolate the AIS data obtained in step S1 to align its time with that of the radar data, and use the second-order spline interpolation method to interpolate the AIS data to obtain the position of the AIS data at the radar scanning time;

[0087] S2.2. Project the radar traces into the parameter space according to the longitude and latitude of the AIS data through the differential transformation method. The expression is:

[0088]

[0089] where P' radar,t is the position of the radar trace after differential transformation at time t, P ais,t is the position of the AIS data at time t, P radar,t is the position of the radar trace at time t, T radar is the time of radar scanning, T ais,start is the time when the first point of the AIS data currently used for matching is located, and T ais,end is the time when the last point of the AIS data currently used for matching is located; it is set that the projected radar trace is represented by (x radar,t , y radar,t ); the result after step processing is as Figure 3 shown;

[0090] S2.3. For the projected radar traces obtained in step S2.2, vote based on the error distribution function according to the Bayesian estimation framework, as shown in the following formula:

[0091] f(x - x radar,t , y - y radar,t )

[0092] Among them, f(x, y) is the error distribution function, and x and y are the variables of the error distribution function;

[0093] The same voting is performed for each projected radar point, and the result is expressed as:

[0094]

[0095] Among them, D t is the set of projected points of all radar point traces in the parameter space at time t, is the counter at time t, used to represent the voting result at time t; The schematic diagram of the voting is as Figure 4 shown;

[0096] Furthermore, when performing step 1 for the first time, assume the error distribution function is:

[0097]

[0098] Among them, f i (lon, lat) is the error distribution function. Since the radar scanning area is large, the error distribution functions of the entire radar scanning area are often different. Therefore, it is necessary to divide the radar scanning area into blocks and number them, and perform independent dynamic optimization on the error distribution functions of different blocks. lon and lat are the longitude and latitude respectively. σ lon , σ lat are the covariances of longitude and latitude respectively, and usually can be obtained by the following formula:

[0099] σ lon = |σ dis ·sinθ|

[0100] Among them, σ dis is the range resolution of the radar, and θ is the angle of the radar's normal north-east. Similarly, σ lat can be expressed as the following formula, with the same parameter meanings:

[0101] σ lat = |σ dis ·cosθ|

[0102] In the second and subsequent iteration processes, the error distribution function will change. In the second and subsequent iteration processes, the error distribution function will change. The error distribution function will be recorded on the hard disk. However, when the program accidentally terminates and starts running, it will still be regarded as the first execution of steps 1-3 to avoid a series of subsequent incorrect matches caused by the early incorrect error distribution function;

[0103] S2.4. For the multiple counters obtained in step S2.3 Use the following filter for filtering:

[0104]

[0105] Among them, H n,t represents the filtering result at time t, n represents (please give the meaning of the letter), and N * represents (please give the meaning of the letter);

[0106] Based on the results of filtering using a threshold of 0.05·n, the results of all filters are accumulated, and the accumulated result is expressed as:

[0107]

[0108] For peak detection is performed, and the deviation of the point with the highest value from the origin is set as the systematic error, and the detected peak is (x peak , y peak ).

[0109] S3. The radar track point indexing module retrieves the radar track points according to the systematic error obtained in step S2, performs a speed correlation check on the retrieved radar track points, proceeds to the next step for the radar track points whose speed parameters meet the requirements, and retrieves the radar track points whose speed parameters do not meet the requirements again until the specified number of times is reached or the radar track points whose speed parameters meet the requirements are retrieved;

[0110] Furthermore, the specific implementation method of step S3 includes the following steps:

[0111] S3.1. Radar track point indexing: For the systematic error obtained in step S2, the following conversion is performed:

[0112] P' ais,t = P ais,t +(x peak , y peak )

[0113] Among them, P' ais,t is the reference point for indexing radar track points at time t;

[0114] The nearest neighbor method is used to select the closest point for each radar track point to obtain the radar track point P radar,t,0 , and the expression is:

[0115]

[0116] Among them, argmin represents finding the point P t in the radar track point set D radar,t that minimizes ||P ais,t - P' radar,t ||2, and || ||2 represents performing here where P radar,t has coordinates (lon radar,t , lat radar,t ), and the corresponding coordinates of P ais,t are (lon ais,t , lat ais,t );

[0117] S3.2. Velocity correlation check: For the radar track point P radar,t,0 output in step S3.1, perform a radial velocity correlation check. If the radial velocity is within three times the radar's velocity resolution range, it is considered that the selection of this radar track point is correct. Output the matching radar track point and execute step S4; otherwise, go to step S3.3.

[0118] S3.3. For the radar track points that do not meet the requirements, remove them from the searched point set and re - execute the following operations:

[0119]

[0120] Then re - perform the velocity correlation check. If it still does not meet the requirements, continue to search further in the set P radar,t ∈D t {P radar,t,0 , P radar,t,1} until the number of repetitions reaches K times. Then stop searching, output all the indexed radar track points, and execute step S4 to obtain the optimal indexed radar track point as P radar,t,max .

[0121] S4. The error function optimization module optimizes the error function based on the PARZEN window for the radar track points with qualified velocity parameters obtained in step S3;

[0122] Furthermore, the specific implementation method of step S4 includes the following steps:

[0123] S4.1. Optimize the error function of the corresponding radar block based on the radar track points indexed in step S3. The expression is:

[0124]

[0125] where D k is all the indexed radar track points located in block i, with a total of k + 1. f i,old is the original error distribution function in block i, and f i,new is the updated error distribution function in block i. (lon P , lat P) is the offset of the indexed radar track point P relative to the systematic error:

[0126] (lon P ,lat P ) = P radar,t,max - P ais,t - (x peak ,y peak );

[0127] S4.2. Re - store the updated error distribution function in step S4.1 into the hard disk in the form of constructing a data table.

[0128] S5. Update the affiliated vessel numbers in the AIS table of the database according to the radar track points whose speed parameters meet the requirements obtained in step S3.

[0129] The method based on Example 1 is illustrated as follows:

[0130] Step S1: Execute initialization:

[0131] Based on mysql5.7, use mysql connector / cpp for management. Create an AIS table for database A in the terminal, and the table attributes are shown in Table 1; at the same time, create a radar point track table for database A, and the table attributes are shown in Table 3.

[0132] During the initialization process, data will also be pre - stored in the AIS table. Assume the data is shown in Table 1.

[0133] Table 3 Data pre - set during the initialization of the AIS table

[0134]

[0135] At the same time, during the initialization process, at least three batches of radar point tracks need to be pre - stored. For the convenience of visualization, only three random points are generated for each batch.

[0136] Table 4 Data pre - fabricated during the initialization of the radar table

[0137] Primary Key 0 1 2 3 4 5 6 7 8 Time 15 15 15 30 30 30 45 45 45 Longitude 0.003 0.0017 0.0 0.001 0.004 0.0032 0.004 0.0055 0.0053 Latitude 0.0014 0.0018 0.0 0.003 0.0037 0.0025 0.0012 0.0057 0.0044 Radial Velocity 2.0 12.7 16.0 3.0 12.9 22.0 11.0 4.0 13.0

[0138] Execute step S1: When the data arrives, determine whether the data is radar data. Assume that the first one is radar point track data and the second one is AIS data, and the two pieces of data arrive simultaneously.

[0139] When the radar data arrives, submit the radar point track data to the corresponding table, and then trigger the signal of the execution program to prepare to execute step S2.

[0140] At the same time, the AIS data also executes the submission task and submits the data entry to the corresponding table.

[0141] Suppose the submitted AIS data is as shown in Table 5:

[0142] Table 5 Newly Submitted AIS Data

[0143]

[0144] The submitted radar data is as shown in Table 6:

[0145] Table 6 Newly Submitted Radar Data

[0146] Primary Key 9 10 11 Time 60 60 60 Longitude 0.0015 0.007 0.0 Latitude 0.007 0.007 0.0043 Radial Velocity 17.0 13.0 13.0

[0147] Execute step S2.

[0148] First, obtain all the MMSI numbers in the database: 000000000, 000000001. Then, for different MMSI numbers, extract all the entries. It is detected that the number of entries corresponding to 00000001 is too small to perform second-order spline interpolation, so it is not executed. For the AIS data corresponding to 000000000, use the interpolation function of the gsl library for interpolation. On the time series of [15, 30, 45, 60], the longitude sequence of the interpolated AIS data is [0.00055, 0.0022, 0.00445, 0.0056], and the latitude sequence is [0.00055, 0.0022, 0.00445, 0.0056]. Assuming the positive north direction is the normal direction of the radar, the speed sequence of the AIS data is [9.19, 9.19, 9.19, 9.05].

[0149] Then, perform a differential transformation on the radar data based on the aligned AIS data. The radar traces in the parameter space after the differential transformation are:

[0150] Table 7 Projected Radar Traces in the Parameter Space

[0151]

[0152] When running for the first time, according to step S2, assume the radar station is at longitude -1 and latitude -1, and the 3σ error of the radar station is 1100m. The variance corresponding to the longitude is The latitude is

[0153] Since there is only one track, assume the block number where all points in this track are located is 0. Thus, according to f0(z, y), that is, f(lon - lon ais , lat - lat ais ) to perform voting, and then the

[0154] Generally, in order to simplify the calculation, a discrete space is used to record Here, it may be assumed that the retrieval longitude is 11m, that is, 0.0001 degrees, so as to obtain 4 counters.

[0155] Use a filter to perform peak detection on the voting space. The maximum value is located at the longitude and latitude [0.001, 0.001]. According to the detected traces, step S3 is executed.

[0156] Execute step S3, radar trace indexing.

[0157] Extract the systematic error according to the detected peak and execute the indexing.

[0158] The nearest points indexed are 1, 4, 7, 10, and the corresponding radial velocities are [12.7, 12.9, 4.0, 13.0] respectively. According to step 3-2, the detection is performed. The radial velocity at 15s should be between [13, 13] (m / s), the velocity at 30s is 13.0m / s, the velocity at 45s should be between [13.1, 13.0] (m / s), and the velocity at 60s should be 12.8m / s. After detection, the velocity of the 7th trace is not within the range. The second closest point is 8. After detection, the velocity of the 8th trace is within the threshold and meets the requirements.

[0159] Output the radar track points matching the 000000000AIS track.

[0160] Execute step four. According to the indexed traces 1, 4, 8, 10, judge the block to which they belong according to the longitude and latitude. Here, it is assumed that the block with the diagonal (0, 0), (0.5, 0.5) is numbered 1, then the corresponding traces are used to optimize the error distribution function f1. According to step S4, a new error distribution function can be obtained as Figure 5 shown.

[0161] When the next batch of radar data arrives, according to the error distribution function in step four, start from step S1 again.

[0162] Embodiment 2:

[0163] A radar trace and AIS track adaptive parallel matching system is realized based on the radar trace and AIS track adaptive parallel matching method described in Embodiment 1, and includes a database, a systematic error extraction module, a radar trace indexing module, and an error function optimization module;

[0164] The database is connected to the systematic error extraction module. The systematic error extraction module is respectively connected to the radar trace indexing module and the error function optimization module. The radar trace indexing module is also connected to the database.

[0165] It should be noted that relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprising", "including" or any other variant thereof are intended to cover non-exclusive inclusion, such that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the phrase "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or device comprising the element.

[0166] Although the present application has been described above with reference to specific embodiments, various improvements can be made thereto and components thereof can be replaced with equivalents without departing from the scope of the present application. In particular, as long as there is no structural conflict, the various features in the specific embodiments disclosed in the present application can be combined with each other in any manner, and the exhaustive description of these combinations is not given in this specification only for the sake of saving space and resources. Therefore, the present application is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. An adaptive parallel matching method for radar point traces and AIS tracks, characterized in that, It includes the following steps: S1. Identify the data arriving at the database. Save the data identified as AIS data into the AIS table of the database and perform data preprocessing. Save the data identified as radar data into the radar table of the database. When the saved radar data exceeds a certain batch, proceed to the next step; S2. The system error extraction module interpolates the AIS data obtained in step S1 to align its time with that of the radar data. Then, project the radar traces into the parameter space according to the longitude and latitude of the AIS data through the differential transformation method. Then, vote on the projected radar traces in the parameter space based on the error distribution function to obtain multiple counters. Filter the obtained multiple counters through a filter. Finally, accumulate the results of all filters and then perform peak detection to extract the system error of the radar traces; S3. The radar trace indexing module retrieves the radar track points based on the system error obtained in step S2, checks the speed correlation of the retrieved radar track points, proceeds to the next step for the radar track points whose speed parameters meet the requirements, and retrieves the radar track points whose speed parameters do not meet the requirements again until the specified number of times is reached or the radar track points whose speed parameters meet the requirements are retrieved; S4. The error function optimization module optimizes the error function based on the PARZEN window for the radar track points whose speed parameters meet the requirements obtained in step S3; S5. Update the affiliated vessel numbers in the AIS table of the database based on the radar track points whose speed parameters meet the requirements obtained in step S3.

2. The adaptive parallel matching method for radar point tracks and AIS tracks according to claim 1, wherein The specific implementation method of step S1 includes the following steps: S1.

1. Set the data extracted from the AIS table of the database to include the Maritime Mobile Service Identity (MMSI), time, longitude, latitude, speed, navigation, and affiliated vessel number. Construct a uniqueness constraint for the composite key of MMSI and time in the AIS data; S1.

2. Set the data extracted from the radar table of the database to include the primary key, time, longitude, latitude, and radial speed; S1.

3. When it is detected that the existing radar data is greater than three batches, proceed to the next step.

3. A method for adaptively and parallelly matching radar dots and AIS tracks according to claim 1 or 2, characterized in that The specific implementation method of step S2 includes the following steps: S2.

1. Interpolate the AIS data obtained in step S1 to align its time with that of the radar data. Use the second-order spline interpolation method to interpolate the AIS data to obtain the position of the AIS data at the radar scan time; S2.

2. Project the radar traces into the parameter space according to the longitude and latitude of the AIS data through the differential transformation method. The expression is: Among them, P' radar,t is the position of the radar track after differential transformation at time t, P ais,t is the position of the AIS data at time t, P radar,t is the position of the radar track at time t, T radar is the time of radar scanning, T ais,start is the time when the first point of the AIS data currently used for matching is located, T ais,end is the time when the last point of the AIS data currently used for matching is located; it is set that the projected radar track is represented by (x radar,t , y radar,t ); S2.

3. For the projected radar traces obtained in step S2.2, vote based on the Bayesian estimation framework and the error distribution function as shown in the following formula: f(x - x radar,t , y - y radar,t ) where f(x,y) is the error distribution function, and x and y are the variables of the error distribution function; Perform the same voting for each projected radar trace, and the result is expressed as: Among them, D t is the set of projection points of all radar traces at time t in the parameter space, is the counter at time t, used to represent the voting result at time t; S2.

4. For the multiple counters obtained in step S2.3 Filter them using the following filter: Among them, H n,t represents the filtering result at time t, n represents the serial number of the filter, and N * represents the set of positive integers; Based on the results of filtering with a threshold of 0.05·n, the results of all filters are accumulated, and the accumulated result is expressed as: For perform peak detection, set the deviation of the point with the highest value from the origin as the systematic error, and obtain the detected peak as (x peak , y peak ).

4. The adaptive parallel matching method for radar dots and AIS tracks according to claim 3, wherein The specific implementation method of step S3 includes the following steps: S3.

1. Radar trace indexing: Perform the following conversion on the system error obtained in step S2: P′ ais,t = P ais,t +(x peak , y peak ) where P' ais,t is the reference point for indexing radar echoes at time t; Select the closest point for each radar plot using the nearest neighbor method to obtain the radar track point P radar,t,0 , and the expression is: where argmin represents finding the point P in the radar point set D t that minimizes ||P radar,t - P' ais,t ||², and || ||² represents taking the second norm of the vector therein; radar,t ​ S3.

2. Velocity correlation check: For the radar track point P output by step S3.1 radar,t,0 , perform a radial velocity correlation check. If the radial velocity is within three times the radar's velocity resolution range, it is considered that the selection of this radar track point is correct. Output the matched radar track point and execute step S4; otherwise, go to step S3.

3. S3.

3. For the radar track points that do not meet the requirements, eliminate such points. For the remaining set D of radar point traces t \{P radar,t,0}, find a new closest point P radar,t,1 , as follows:: Then, perform the speed correlation check again. If it still does not meet the requirements, continue to search further in the set P radar,t ∈D t {P radar,t,0 ,P radar,t,1} until the number of repetitions reaches K times. Then, stop the search, output all the radar track points obtained by indexing, and execute step S4 to obtain the optimal radar track point obtained by indexing as P radar,t,max .

5. The adaptive parallel matching method for radar point traces and AIS tracks according to claim 4, wherein The specific implementation method of step S4 includes the following steps: S4.

1. Optimize the error function of the corresponding radar block based on the radar track points indexed in step S3. The expression is as follows: Among them, D k is all the indexed radar track points located in block i, and it is assumed that there are k + 1 in total. f i,old is the old error distribution function in block i, and f i,new is the updated error distribution function in block i. (lon P , lat P ) is the offset of the indexed radar track point P relative to the systematic error: (lon P , lat P ) = P radar,t,max -P ais,t -(x peak , y peak ); S4.

2. Reposit the updated error distribution function in step S4.1 into the hard disk in the form of constructing a data table.

6. A radar point track and AIS track adaptive parallel matching system, which is realized relying on the radar point track and AIS track adaptive parallel matching method described in any one of claims 1-5, is characterized in that It includes a database, a system error extraction module, a radar point track indexing module, and an error function optimization module; The database is connected to the system error extraction module. The system error extraction module is respectively connected to the radar point track indexing module and the error function optimization module. The radar point track indexing module is also connected to the database.

Images