A track matching method for sky wave radar sea detection based on multi-source data

Abstract

The application relates to the field of radar positioning precision analysis, and discloses a track matching method for sea detection of sky wave radar based on multi-source data, which comprises the following steps: original data acquisition and multi-stage preprocessing, redundant information filtering based on detection characteristics, multi-domain track matching parameter configuration, track matching based on multi-feature parameters, and comprehensive evaluation of track matching results. The application improves the accuracy and availability of track matching.

Description

Technical Field

[0001] This application relates to the field of radar positioning accuracy analysis technology, and in particular to a track matching method for sea detection by skywave radar based on multi-source data. Background Technology

[0002] The performance evaluation, positioning accuracy analysis, and positioning enhancement of radar for maritime detection all rely on accurate track matching results. Radar track association matching based on high-precision location data of sea surface targets, such as AIS and GPS, has been widely used in radar data analysis systems. As an effective device for over-the-horizon detection and early warning of the sea surface, skywave radar can continuously monitor targets thousands of kilometers away by reflecting high-frequency electromagnetic waves through the ionosphere. However, the complex characteristics of maritime detection mean that track matching with high-precision target location data requires consideration of various factors.

[0003] Recent research and patent applications have disclosed methods for radar track matching using AIS. For example, patent application CN202510507458.4 describes an adaptive parallel matching method and system for radar points and AIS tracks. This method completes the matching of radar points and AIS tracks by voting and filtering the projection of radar points, considering the accumulated peak system error and velocity correlation. Patent application CN202510821286.8 describes a real-time matching method and system for radar and AIS multi-source tracks based on serialization decision-making. This method constructs radar-AIS initial screening track pairs, sets a sliding time window to obtain track segments, and uses these segments to construct a track matching neural network to obtain the time series matching probability. Patent application CN202411302814.0 describes a radar and AIS track association method, device, and equipment based on spatiotemporal attention. This method extracts local spatial and temporal features of track data, uses a cross-attention mechanism to calculate the association probability matrix of radar and AIS tracks, and provides matrix parameters for track association matching.

[0004] The above methods can achieve the expected results when targeting radar systems with high detection accuracy, but they are not suitable for skywave radars with low detection accuracy, long range, and a lot of interference. They lack consideration for factors such as detection characteristics, environment, and data quality, and it is difficult to obtain reliable training samples by simply relying on data pairing. Summary of the Invention

[0005] In view of this, this application provides a track matching method for sea detection by skywave radar based on multi-source data, which improves the accuracy and usability of track matching.

[0006] This application discloses a track matching method for sea-penetrating radar based on multi-source data, which includes: Step 1: Raw data acquisition and multi-level preprocessing, mainly completing multi-source data aggregation, core element extraction, abnormal data removal, data consistency processing, and data structuring processing; Step 2: Redundant information filtering based on detection characteristics, mainly to filter targets with a speed less than a preset speed threshold, targets with a radar cross section less than a preset cross section threshold, spatiotemporal misalignment information, and track information; Step 3: Configure multi-domain track matching parameters, mainly including configuring spatiotemporal domain matching parameters, energy domain matching parameters, and characteristic parameter matching parameters; Step 4: Track matching based on multiple feature parameters, mainly completing track matching based on spatiotemporal energy, Bragg interference removal, spatial matching parameter correction, radial velocity and longest track determination; Step 5: Comprehensive evaluation of track matching results. This mainly involves calculating evaluation indicators, including average radial velocity jitter, average single track matching rate, and overall matching rate.

[0007] Further, step 1 includes: Step 1.1: Read and parse various types of data from multi-source data and classify and store them; multi-source data includes files, data streams, databases, and service interfaces; various types of data include skywave radar position data (i.e., radar position), skywave radar detection data (i.e., detection data), target precise position data (i.e., high-precision data), target static size parameters (i.e., target size), and meteorological and hydrological environmental parameters (i.e., environmental parameters); among them, high-precision data includes AIS data of civilian ships and Beidou / GPS positioning data of experimental ships; Step 1.2: Filter the valid information in various types of data, including: Radar location: retain radar longitude (Lonr) and radar latitude (Latr); Detection data: retain target detection time, target detection batch number, target detection longitude, target detection latitude, target detection radial velocity, radar transmission frequency, target detection range, target detection azimuth, and target echo amplitude; detection data corresponding to the same target detection batch number constitute a detection track; High-precision data: retain the time, unique code, precise longitude, precise latitude, precise velocity, and precise heading corresponding to the target's precise location data; high-precision data corresponding to the same unique code constitute a high-precision track; Size parameters: retain MMSI code, ship length, ship width, and ship height; Environmental parameters: retain current time, longitude of current location, latitude of current location, and sea wave height. Step 1.3: Remove data from the filtered data in Step 1.2 that has missing features, contains null values, has invalid values, or contains duplicates; Step 1.4: Perform dimension and precision consistency processing on the data after removing abnormal data in Step 1.3; wherein, time is converted to Epoch time, longitude, latitude, frequency and wave height are converted to numerical values, unique code and detection batch number are dimensionless integers, AIS data uses MMSI code, and the unique code of other ships is equal to the current maximum code value plus 1. Step 1.5: Perform single-target clustering on the detection data and high-precision data respectively. Classify the detection data according to batch number, and then sort the data within each batch number in ascending order of time. Classify the high-precision data according to unique code, and then sort the data within each batch number in ascending order of time.

[0008] Further, step 2 includes: Step 2.1: Based on the detection characteristics, exclude irrelevant targets with radial velocities less than the preset velocity threshold, that is, filter the target data with precise velocities less than the velocity threshold in the high-precision data obtained in Step 1.5; Step 2.2: Obtain the minimum scattering cross section (RCS) for stable sea-view detection; load the dimensional parameters processed in Step 1.4 and the high-precision data processed in Step 2.1, and calculate the RCS of each ship's side view by associating them through unique codes. Where A is the side area of ​​the ship and λ is the radar wavelength, high-precision data with RCS smaller than the minimum scattering cross section are filtered out. Step 2.3: Exclude probe data and high-precision data that are not within the same spatiotemporal range: Based on the results of Step 2.2, assume that the spatiotemporal distribution range of the probe data is... , , The spatiotemporal distribution range of high-precision data is , , Then filter the time in high-precision data In addition, longitude in and latitude Other data; filtering the time in the detection track In addition, longitude in In addition, and latitude Other data; , These represent the minimum and maximum temporal values ​​of the spatiotemporal distribution of the probe data, respectively. , These represent the minimum and maximum latitudinal values ​​of the spatiotemporal distribution of the probe data, respectively. , These represent the minimum and maximum longitude values ​​of the spatiotemporal distribution of the probe data, respectively. , These represent the minimum and maximum temporal values ​​of the spatiotemporal distribution of high-precision data, respectively. , These represent the minimum and maximum latitudinal values ​​of the spatiotemporal distribution of high-precision data, respectively. , These represent the minimum and maximum longitude values ​​of the spatiotemporal distribution of high-precision data, respectively. This indicates the preset maximum tolerance range for spatial latitude and longitude. Step 2.4: Under stable and continuous detection conditions, perform matching to filter detection data and high-precision data in a single track where the number of points is less than the preset number.

[0009] Further, step 3 includes: Step 3.1: Obtain the revisit period of the skywave radar's sea detection. Distance accuracy Azimuth accuracy The design time matching parameters are: , The maximum time tolerance factor; obtain ground latitude. corresponding distance Assuming there are y illumination areas, the spatial matching parameters for detection data falling within a certain illumination area are designed as follows:

[0010] in, Indicates the longitude matching parameter. Indicates the latitude matching parameter. This indicates the distance from the center point of the illuminated area to the radar; Step 3.2: Determine the echo energy of a single detection track, assuming the median detection amplitude is... amplitude at each point and The relative deviation is The quality level of the data is then determined based on the relative deviation; the quality levels include A, B, and C; the relative deviation corresponding to level B is between that of level A and level C. Step 3.3: Selecting radial velocity as the characteristic parameter, the characteristic parameter matching parameters for radial velocity are configured as follows:

[0011] in, Characteristic matching parameters representing radial velocity. This indicates the radial velocity of the target detection.

[0012] Further, step 4 includes: Step 4.1: Perform threshold matching between each probe track and each high-precision track; extract a probe track and determine its median amplitude. Calculate the amplitude of each data point and the median of the amplitude. If the deviation is found, data of grade C is removed based on the quality level; if the detected data is grade A or B, all high-precision tracks that overlap with the time of the track are obtained, and the data in each high-precision track is judged one by one according to the spatial matching parameters based on the high-precision track as the benchmark; if the number of points of a certain detection track or a certain high-precision track after matching is less than the preset value, the detection track or high-precision track is removed. Step 4.2: Based on the processing results of Step 4.1, eliminate the influence of the first-order Bragg peak and the broadening of the Doppler velocity of sea clutter; calculate the radial velocity corresponding to the first-order Bragg peak. Where g is the acceleration due to gravity, the actual Bragg clutter bandwidth and its corresponding radial velocity are calculated using sea state data. ,in, For directional distribution and nonlinear engineering fitting coefficients, For Bragg frequency, For steep waves, H s For the high waves, If the dominant wavelength of the ocean wave is used, then the radial velocity of the disturbance is... ; Remove radial velocities from each batch of probe tracks in step 4.1 that fall into the range. Data within the specified range; tracks in the matching data whose number of detection tracks or high-precision track points is less than a preset value are removed; Step 4.3: Based on the results obtained in Step 4.2, select a set of matched data, assuming the detection batch number is... ,get ,like In the spatial parameters of latitude and longitude Updated to ;like In the spatial parameters of latitude and longitude Updated to ;like or If there are changes, repeat steps 3.1, 4.1, 4.2, and 4.3 until... and All converge to 1; This is the azimuth accuracy reduction factor. This is the distance accuracy contraction factor; Step 4.4: Calculate the radial velocity component of the high-precision data relative to the skywave radar. ,in, , , The target's azimuth angle relative to the radar. The heading angle of the target. For precise speed, The velocity component is the velocity component moving eastward from the target. For the northward velocity component of the target, extract the matching data filtered in step 4.2, if the detection batch number... For each high-precision track, the decision parameters are calculated based on multiple unique codes. Keep the largest The corresponding high-precision flight path.

[0013] Furthermore, in step 4.1, after thresholding the high-precision data, the matching points within a certain irradiation area conform to the following relationship:

[0014] in, , These are two adjacent time points, representing a detection track and a high-precision track, respectively. , These are the probe tracks The latitude and longitude corresponding to a given time. , These are the high-precision flight paths. The latitude and longitude corresponding to the time.

[0015] Furthermore, in step 4.3 and The calculation formula is:

[0016]

[0017]

[0018]

[0019]

[0020] in, This is the set of probe track vectors corresponding to the matched data after processing in step 4.2. To meet the distance determination criteria The set of detection track vectors, Θ i To meet the conditions for determining direction The set of track vectors, , They are respectively The detection distance and detection orientation at any given moment. , They are respectively The high-precision data position relative to the skywave radar at any given time, including distance and azimuth. It is the root mean square. for High-precision data of the radial velocity relative to the skywave radar at all times.

[0021] Furthermore, in step 4.4, the formula for calculating the decision parameter is as follows:

[0022] in, l v To meet the requirements of this high-precision trajectory The number of waypoints , These represent the radial velocities at a set of adjacent moments for the detection track and the high-precision track, respectively. For detection batch number The number of waypoints corresponding to the longest high-precision track. This represents the number of waypoints for this high-precision track.

[0023] Further, step 5 includes: Step 5.1: Obtain the total number of track points detected before matching. Total number of all matched probe tracks and the total number of corresponding waypoints. ; Step 5.2: Calculate the average radial velocity jitter for all matched data. Average single-track matching rate Overall matching rate .

[0024] Furthermore, in step 5.2, The calculation formula is:

[0025] in, This represents the number of trackpoints for a given matched probe track. This is the first of the exploration tracks. The difference between a radial velocity and the radial velocity at the nearest neighbor time point of the corresponding high-precision track. This is the average of all radial velocity differences along the probe track; The calculation formula is:

[0026] in, For the detection batch number The number of track points matched by the detected track. To match the total number of track points of the previous probe track; The calculation formula is: .

[0027] Due to the adoption of the above technical solution, this application has the following advantages: 1. This invention fully considers the complexity of skywave radar's sea detection, introduces data from multiple sources and dimensions, and designs a hierarchical matching strategy, which can greatly improve the accuracy of skywave radar in matching the tracks of various sea surface targets under complex and real sea conditions.

[0028] 2. This invention utilizes expert knowledge, index parameters, and data characteristics to design convergence rules to train the matching threshold in the space, and designs decision parameters using long tracks and quasi-radial velocities to avoid "one-to-many" situations and ensure the uniqueness of the matching results.

[0029] 3. This invention combines the sea-penetrating characteristics of skywave radar and sets up multiple filtering rules. By filtering core fields and eliminating redundant information, the complexity of subsequent calculations can be effectively avoided and the operating efficiency can be improved.

[0030] 4. This invention proposes a comprehensive evaluation index for track matching, which can quantify the matching effect and has reference value for further equipment performance evaluation and positioning accuracy analysis. Attached Figure Description

[0031] 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 recorded in the embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings.

[0032] Figure 1 The diagram shows a schematic flowchart of a method for track matching in sea detection using skywave radar based on multi-source data, as proposed in an embodiment of the present invention.

[0033] Figure 2 A schematic diagram of the multi-domain track matching parameter configuration calculation process proposed in an embodiment of the present invention is shown.

[0034] Figure 3 A schematic diagram of the multi-feature parameter track matching rule process proposed in an embodiment of the present invention is shown. Detailed Implementation

[0035] The present application will be further described in conjunction with the accompanying drawings and embodiments. The described embodiments are only some, not all, of the embodiments of the present application. All other embodiments obtained by those skilled in the art should fall within the protection scope of the embodiments of the present application.

[0036] See Figures 1 to 3 This application provides an embodiment of a track matching method for sea detection using skywave radar based on multi-source data, comprising the following steps: Step 1: Raw data acquisition and multi-level preprocessing, mainly completing multi-source data aggregation, core element extraction, abnormal data removal, data consistency processing, and data structuring.

[0037] Step 2: Redundant information filtering based on detection characteristics, mainly to filter low-speed targets (targets with a speed less than a preset speed threshold), small-scattering targets (targets with a radar cross section less than a preset cross section threshold), spatiotemporal misalignment information, and short-track information.

[0038] Step 3: Configure multi-domain track matching parameters, mainly including configuring spatiotemporal domain matching parameters, energy domain matching parameters, and characteristic parameter matching parameters.

[0039] Step 4: Track matching based on multiple feature parameters, mainly completing track matching based on spatiotemporal energy, Bragg interference removal, spatial matching parameter correction, radial velocity and longest track determination.

[0040] Step 5: Comprehensive evaluation of track matching results, mainly involving the calculation of evaluation indicators such as average radial velocity jitter, average single track matching rate, and overall matching rate.

[0041] In step 1, multi-source data aggregation is performed on the raw data, and on this basis, multi-level preprocessing is carried out, including core element extraction, outlier removal, data consistency processing, and data structuring. The specific implementation method includes the following steps: Step 1.1: Conduct multi-source data aggregation, reading and parsing various types of data from data files, data streams, databases, and service interfaces, including five categories: radar position, detection data, high-precision data, target size, and environmental parameters, and storing them accordingly. This embodiment processes the data files, obtaining five data tables based on the five required data categories. The high-precision data includes AIS data from civilian vessels and BeiDou / GPS positioning data from the experimental vessel.

[0042] Step 1.2: Extract core elements and filter the valid information from the above 5 data tables. Specifically: Radar location: retain radar longitude and latitude; Detection data: retain target detection time, target detection batch number, target detection longitude, target detection latitude, target detection radial velocity, radar transmission frequency, target detection range, target detection azimuth, and target echo amplitude; High-precision data: retain the time, unique code, precise longitude, precise latitude, precise velocity, and precise heading corresponding to the target's precise location data; Size parameters: retain MMSI code, ship length, ship width, and ship height; Environmental parameters: retain current time, longitude, latitude, and sea wave height (Note: All elements in each valid data entry of the above 5 categories must exist simultaneously; otherwise, the data entry is considered invalid). Element information is shown in Table 1: Table 1 Element Information

[0043] Step 1.3 involves anomaly data removal. Data incomplete or missing elements from the five data categories in Step 1.2 are removed. Invalid values ​​are also removed, such as data containing one or more of the following: |Longitude| > 180°, |Latitude| > 90°, ship length > 500m, ship width > 150m, wave height < 0. Duplicate values ​​are removed, such as: ① multiple data entries from the same time and batch number for the probe data; ② multiple data entries from the same time and unique code for the high-precision data; ③ multiple data entries from the same MMSI code for the size parameters; ④ multiple data entries from the same time and latitude / longitude for the environmental data. In these four cases, only the first data entry is retained. Tracks with fewer than 8 single target track points in the probe and high-precision data are deleted. In this embodiment, the results of anomaly data removal are shown in Table 2. Table 2 Results after outlier removal

[0044] Step 1.4 involves data consistency processing, specifically performing dimension and precision consistency processing on the five types of data from Step 1.3. Specifically, time is converted to Epoch time in seconds (integer); longitude and latitude are converted to numerical values ​​in degrees (6 decimal places); frequency is converted to numerical values ​​in megahertz (2 decimal places); unique codes are dimensionless integers; AIS data uses MMSI codes, and codes for other high-precision ships can be obtained by incrementing the maximum MMSI value by 1; target detection batch numbers are dimensionless integers; and wave height is converted to numerical values ​​in meters (1 decimal place).

[0045] Step 1.5: Perform data structuring processing. Perform single-target stacking on the detection data and high-precision data respectively. Classify the detection data according to batch number, and then sort the data in each batch number in ascending order of time. Classify the high-precision data according to unique code, and then sort the data in each unique code in ascending order of time.

[0046] In step 2, redundant information filtering is performed based on the above preprocessed data. The specific implementation method includes the following steps: Step 2.1: Perform low-slow target filtering. Load all pre-processed high-precision data, and based on detection characteristics, exclude irrelevant targets with radial velocities less than a preset radial velocity. Directly filter targets with velocities less than the precision velocities in the high-precision data from Step 1.5. In this embodiment, the results of low-slow target filtering are shown in Table 3: Table 3 Results after filtering slow and slow targets

[0047] Step 2.2: Perform small scattering target filtering, consult the technical manual to obtain the minimum scattering cross section (RCSmin) for stable sea-based detection; load the AIS high-precision data and size parameters processed in Step 2.1, and calculate the side-view RCS for each ship by associating with the MMSI code. (Where A is the lateral surface area of ​​the ship,) The radar wavelength can be obtained from the radar frequency in the detection data. , here (Taking the minimum value), high-precision data with RCS less than RCSmin are filtered. In this embodiment, the results of filtering small scattering targets are shown in Table 4: Table 4 Results after filtering small scattering targets

[0048] Step 2.3: Perform spatiotemporal misalignment information filtering to exclude detection data and high-precision data that are not within the same spatiotemporal range. Based on the results of Step 2.2, assume the spatiotemporal distribution range of the detection data is... , , The spatiotemporal distribution range of high-precision data is , , Then filter the time in high-precision data In addition, longitude in and latitude Other data; filtering the time in the detection track In addition, longitude in In addition, and latitude Other data; , These represent the minimum and maximum temporal values ​​of the spatiotemporal distribution of the probe data, respectively. , These represent the minimum and maximum latitudinal values ​​of the spatiotemporal distribution of the probe data, respectively. , These represent the minimum and maximum longitude values ​​of the spatiotemporal distribution of the probe data, respectively. , These represent the minimum and maximum temporal values ​​of the spatiotemporal distribution of high-precision data, respectively. , These represent the minimum and maximum latitudinal values ​​of the spatiotemporal distribution of high-precision data, respectively. , These represent the minimum and maximum longitude values ​​of the spatiotemporal distribution of high-precision data, respectively. In this embodiment, the results after filtering the spatiotemporal misalignment information are shown in Table 5: Table 5 Results after filtering spatiotemporal misalignment information

[0049] Step 2.4: Perform short track information filtering. To improve the accuracy of matching results, matching should be performed under stable and continuous detection conditions, filtering out detection and high-precision data with fewer than 8 points in a single track. In this embodiment, the results after short track information filtering are shown in Table 6: Table 6 Results after filtering short flight track information

[0050] In step 3, the multi-domain track matching parameters are configured. Matching parameters for the time domain, spatial domain, energy domain, and characteristic parameters are calculated using technical manuals and expert knowledge to provide judgment boundaries for subsequent track matching. The specific implementation method includes the following steps: Step 3.1: Obtain the revisit period T0 and range accuracy of the skywave radar for sea detection. Azimuth accuracy The design time matching parameters are: ; Obtain the distance corresponding to 1° latitude on the ground. Assuming there are y illumination areas, the spatial matching parameters for detection data falling within a certain illumination area are designed as follows:

[0051] in, Indicates the longitude matching parameter. Indicates the latitude matching parameter. This indicates the distance from the center point of the illumination area to the radar.

[0052] In this embodiment, the configuration results of the spatiotemporal domain matching parameters are shown in Table 7: Table 7. Spatiotemporal Domain Matching Parameter Configuration Results

[0053] Step 3.2: Configure energy domain matching parameters. To improve the quality of the detection data, determine the echo energy of a single track after spatiotemporal matching. Assume the median detection amplitude is... amplitude at each point and The relative deviation is The data quality level is determined based on the relative amplitude deviation. In this embodiment, the energy domain matching parameters are configured as shown in Table 8: Table 8 Energy Domain Matching Parameter Configuration

[0054] Step 3.3: Configure characteristic parameter matching parameters. Selecting radial velocity as the characteristic parameter, the radial velocity characteristic parameter matching parameters are configured as follows:

[0055] in, Characteristic matching parameters representing radial velocity. This indicates the radial velocity of the target detection.

[0056] In this embodiment, the radial velocity characteristic parameter matching parameter is set to 3 m / s.

[0057] In step 4, track matching based on multiple feature parameters is performed. Based on the filtered data and various matching parameters described above, multi-level matching of tracks across multiple dimensions is conducted. The specific implementation method includes the following steps: Step 4.1: Conduct spacetime energy-based track matching, matching probe data with high-precision data using a "multiple-to-multiple" approach. Different tracks are distinguished by probe data batch numbers, and threshold matching is performed one-by-one with the target tracks under the unique target codes in the high-precision data. Specifically, the track with batch number... p 1i The median amplitude of the detected data was obtained. Calculate the amplitude at each point and Deviation, according to the quality grade obtained from Table 8 in step 3.2, results with grade C are discarded. If For level A or B, acquire all high-precision data that overlaps with the time of the flight track. For each high-precision flight path, a threshold judgment is performed. Matching points within a certain illumination area should meet the following relationship:

[0058] in, , These are two adjacent time points, representing a detection track and a high-precision track, respectively. , These are the probe tracks The latitude and longitude corresponding to a given time. , These are the high-precision flight paths. The latitude and longitude corresponding to the time.

[0059] If the number of points on a certain detection track or high-precision track is less than 5 after spatiotemporal matching, it will be discarded.

[0060] In this embodiment, the trajectory matching results based on spatiotemporal energy are shown in Table 9: Table 9. Track matching results based on spatiotemporal energy

[0061] Step 4.2: Perform Bragg interference removal. Based on the results of Step 4.1, further eliminate the influence of the first-order Bragg peak and sea clutter Doppler velocity broadening. Calculate the radial velocity corresponding to the first-order Bragg peak. (where g is taken as) ); Calculate the actual Bragg clutter bandwidth using sea state data, and the corresponding radial velocity of the interference. (in, Then the radial velocity of the disturbance Perform radial velocity range determination, and eliminate data from each batch of detection data in step 4.1 whose radial velocity falls within the range. Data filtering within the range. Furthermore, cases where the number of detected tracks or high-precision track points is less than 5 are removed from all matched data.

[0062] In this embodiment, the Bragg interference removal results are shown in Table 10: Table 10 Results of Bragg interference removal

[0063] Step 4.3: Iterate the latitude and longitude matching parameters. To avoid the "one-to-many" situation caused by excessively large spatiotemporal parameters in 4.1, perform a shrinking process. Take the results from Step 4.2, aggregate the information, and perform a regression analysis on the probe track. p 1i High-precision flight paths p 2j Calculate the shrinkage coefficient:

[0064]

[0065]

[0066]

[0067]

[0068] in, This is the set of probe track vectors corresponding to the matched data after processing in step 4.2. To meet the distance determination criteria The set of detection track vectors, Θ i To meet the conditions for determining direction The set of track vectors, , They are respectively The detection distance and detection orientation at any given moment. , They are respectively The high-precision data position relative to the skywave radar at any given time, including distance and azimuth. It is the root mean square. for High-precision data of the radial velocity relative to the skywave radar at all times.

[0069] like In the spatial parameters of latitude and longitude Updated to ;like In the spatial parameters of latitude and longitude Updated to .like or If there are changes, repeat steps 3.1, 4.1, 4.2, and 4.3 until... and All converge to 1.

[0070] In this embodiment, after 3 cycles, the two contraction coefficients converge to 1.

[0071] Step 4.4: Conduct a uniqueness determination based on radial velocity and the longest trajectory to further address potential "one-to-many" problems. Calculate the radial velocity component of the target's high-precision coordinates relative to the skywave radar. (in, , , The target's azimuth angle relative to the radar. (Target high-precision heading angle). Obtain the filtered matching data from step 4.2. If the detection batch number p 1i For each of the multiple unique codes, the decision parameters for each high-precision track are calculated. Keep the largest The corresponding high-precision track is calculated using the following decision parameters:

[0072] in, lv To meet the requirements of this high-precision trajectory The number of waypoints , These represent the radial velocities at a set of adjacent moments for the detection track and the high-precision track, respectively. For detection batch number The number of waypoints corresponding to the longest high-precision track. This represents the number of waypoints for this high-precision track.

[0073] Keep the largest The corresponding high-precision flight path.

[0074] In this embodiment, the radial velocity and the longest trajectory determination results are shown in Table 11: Table 11. Decision Results of Radial Velocity and Longest Track

[0075] In step 5, based on the matching results above, various evaluation indicators are calculated by comparing them with the original data. The specific implementation method includes the following steps: Step 5.1: Obtain the total number of track points detected before matching. Get the total number of all matched probe tracks. and the total number of corresponding waypoints. In this embodiment, =8010, =45, =1600.

[0076] Step 5.2: Calculate the average radial velocity jitter for all matched data. Average single-track matching rate Overall matching rate In this embodiment, =1.485, =22.05%, = 19.98%.

[0077] In step 5.2, The calculation formula is:

[0078] in, This represents the number of trackpoints for a given matched probe track. This is the first of the exploration tracks. The difference between a radial velocity and the radial velocity at the nearest neighbor time point of the corresponding high-precision track. This is the average of all radial velocity differences along the probe track; The calculation formula is:

[0079] in, For the detection batch number The number of track points matched by the detected track. To match the total number of track points of the previous probe track; The calculation formula is: .

[0080] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application and not to limit them. Although this application has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of this application. Any modifications or equivalent substitutions that do not depart from the spirit and scope of this application should be covered within the protection scope of the claims of this application.

Claims

1. A method for track matching in sea-based skywave radar detection using multi-source data, characterized in that, include: Step 1: Raw data acquisition and multi-level preprocessing, mainly completing multi-source data aggregation, core element extraction, abnormal data removal, data consistency processing, and data structuring processing; Step 2: Redundant information filtering based on detection characteristics, mainly to filter targets with a speed less than a preset speed threshold, targets with a radar cross section less than a preset cross section threshold, spatiotemporal misalignment information, and track information; Step 3: Configure multi-domain track matching parameters, mainly including configuring spatiotemporal domain matching parameters, energy domain matching parameters, and characteristic parameter matching parameters; Step 4: Track matching based on multiple feature parameters, mainly completing track matching based on spatiotemporal energy, Bragg interference removal, spatial matching parameter correction, radial velocity and longest track determination; Step 5: Comprehensive evaluation of track matching results. This mainly involves calculating evaluation indicators, including average radial velocity jitter, average single track matching rate, and overall matching rate.

2. The method according to claim 1, characterized in that, Step 1 includes: Step 1.1: Read and parse various types of data from multi-source data and classify and store them; multi-source data includes files, data streams, databases, and service interfaces; various types of data include skywave radar position data (i.e., radar position), skywave radar detection data (i.e., detection data), target precise position data (i.e., high-precision data), target static size parameters (i.e., target size), and meteorological and hydrological environmental parameters (i.e., environmental parameters); among them, high-precision data includes AIS data of civilian ships and Beidou / GPS positioning data of experimental ships; Step 1.2: Filter the valid information in various types of data, including: Radar location: retain radar longitude (Lonr) and radar latitude (Latr); Detection data: retain target detection time, target detection batch number, target detection longitude, target detection latitude, target detection radial velocity, radar transmission frequency, target detection range, target detection azimuth, and target echo amplitude; detection data corresponding to the same target detection batch number constitute a detection track; High-precision data: retain the time, unique code, precise longitude, precise latitude, precise velocity, and precise heading corresponding to the target's precise location data; high-precision data corresponding to the same unique code constitute a high-precision track; Size parameters: retain MMSI code, ship length, ship width, and ship height; Environmental parameters: retain current time, longitude of current location, latitude of current location, and sea wave height. Step 1.3: Remove data from the filtered data in Step 1.2 that has missing features, contains null values, has invalid values, or has duplicates; Step 1.4: Perform dimension and precision consistency processing on the data after removing abnormal data in Step 1.3; wherein, time is converted to Epoch time, longitude, latitude, frequency and wave height are converted to numerical values, unique code and detection batch number are dimensionless integers, AIS data uses MMSI code, and the unique code of other ships is equal to the current maximum code value plus 1. Step 1.5: Perform single-target clustering on the detection data and high-precision data respectively. Classify the detection data according to batch number, and then sort the data in each batch number in ascending order of time. Classify the high-precision data according to unique code, and then sort the data in each batch number in ascending order of time.

3. The method according to claim 2, characterized in that, Step 2 includes: Step 2.1: Based on the detection characteristics, exclude irrelevant targets with radial velocities less than the preset velocity threshold, that is, filter the target data with precise velocities less than the velocity threshold in the high-precision data obtained in Step 1.5; Step 2.2: Obtain the minimum scattering cross section (RCS) for stable sea-view detection; load the dimensional parameters processed in Step 1.4 and the high-precision data processed in Step 2.1, and calculate the RCS of each ship's side view by associating them through unique codes. Where A is the side area of ​​the ship and λ is the radar wavelength, high-precision data with RCS smaller than the minimum scattering cross section are filtered out. Step 2.3: Exclude probe data and high-precision data that are not within the same spatiotemporal range: Based on the results of Step 2.2, assume that the spatiotemporal distribution range of the probe data is... , , The spatiotemporal distribution range of high-precision data is , , Then filter the time in high-precision data In addition, longitude in and latitude Other data; filtering the time in the detection track In addition, longitude in In addition, and latitude Other data; , These represent the minimum and maximum temporal values ​​of the spatiotemporal distribution of the probe data, respectively. , These represent the minimum and maximum latitudinal values ​​of the spatiotemporal distribution of the probe data, respectively. , These represent the minimum and maximum longitude values ​​of the spatiotemporal distribution of the probe data, respectively. , These represent the minimum and maximum temporal values ​​of the spatiotemporal distribution of high-precision data, respectively. , These represent the minimum and maximum latitudinal values ​​of the spatiotemporal distribution of high-precision data, respectively. , These represent the minimum and maximum longitude values ​​of the spatiotemporal distribution of high-precision data, respectively. This indicates the preset maximum tolerance range for spatial latitude and longitude. Step 2.4: Under stable and continuous detection conditions, perform matching to filter detection data and high-precision data in a single track where the number of points is less than the preset number.

4. The method according to claim 1, characterized in that, Step 3 includes: Step 3.1: Obtain the revisit period of the skywave radar's sea detection. Distance accuracy Azimuth accuracy The design time matching parameters are: , The maximum time tolerance factor; obtain ground latitude. corresponding distance Assuming there are y illumination areas, the spatial matching parameters for detection data falling within a certain illumination area are designed as follows: in, Indicates the longitude matching parameter. Indicates the latitude matching parameter. This indicates the distance from the center point of the illuminated area to the radar; Step 3.2: Determine the echo energy of a single detection track, assuming the median detection amplitude is... amplitude at each point and The relative deviation is The quality level of the data is then determined based on the relative deviation; the quality levels include A, B, and C; the relative deviation corresponding to level B is between that of level A and level C. Step 3.3: Selecting radial velocity as the characteristic parameter, the characteristic parameter matching parameters for radial velocity are configured as follows: in, Characteristic matching parameters representing radial velocity. This indicates the radial velocity of the target detection.

5. The method according to claim 1, characterized in that, Step 4 includes: Step 4.1: Perform threshold matching between each probe track and each high-precision track; extract a probe track and determine its median amplitude. Calculate the amplitude of each data point and the median of the amplitude. If the deviation is found, data of grade C is removed based on the quality level; if the detected data is grade A or B, all high-precision tracks that overlap with the time of the track are obtained, and the data in each high-precision track is judged one by one according to the spatial matching parameters based on the high-precision track as the benchmark; if the number of points of a certain detection track or a certain high-precision track after matching is less than the preset value, the detection track or high-precision track is removed. Step 4.2: Based on the processing results of Step 4.1, eliminate the influence of the first-order Bragg peak and the broadening of the Doppler velocity of sea clutter; calculate the radial velocity corresponding to the first-order Bragg peak. Where g is the acceleration due to gravity, the actual Bragg clutter bandwidth and its corresponding radial velocity are calculated using sea state data. ,in, For directional distribution and nonlinear engineering fitting coefficients, For Bragg frequency, For steep waves, H s For the high waves, If the dominant wavelength of the ocean wave is used, then the radial velocity of the disturbance is... ; Remove radial velocities from each batch of probe tracks in step 4.1 that fall into the range. Data within the specified range; tracks in the matching data whose number of detection tracks or high-precision track points is less than a preset value are removed; Step 4.3: Based on the results obtained in Step 4.2, select a set of matched data, assuming the detection batch number is... ,get ,like In the spatial parameters of latitude and longitude Updated to ;like In the spatial parameters of latitude and longitude Updated to ;like or If there are changes, repeat steps 3.1, 4.1, 4.2, and 4.3 until... and All converge to 1; This is the azimuth accuracy reduction factor. This is the distance accuracy contraction factor; Step 4.4: Calculate the radial velocity component of the high-precision data relative to the skywave radar. ,in, , , The target's azimuth angle relative to the radar. The heading angle of the target. For precise speed, The velocity component is the velocity component heading eastward from the target. For the northward velocity component of the target, extract the matching data filtered in step 4.2, if the detection batch number... For each high-precision track, the decision parameters are calculated based on multiple unique codes. Keep the largest The corresponding high-precision flight path.

6. The method according to claim 5, characterized in that, In step 4.1, after thresholding the high-precision data, the matching points within a certain irradiation area meet the following relationship: in, , These are two adjacent time points, representing a detection track and a high-precision track, respectively. , These are the probe tracks The latitude and longitude corresponding to the time. , These are the high-precision flight paths. The latitude and longitude corresponding to the time.

7. The method according to claim 5, characterized in that, In step 4.3 and The calculation formula is: in, This is the set of probe track vectors corresponding to the matched data after processing in step 4.

2. To meet the distance determination criteria The set of detection track vectors, Θ i To meet the conditions for determining direction The set of track vectors, , They are respectively The detection distance and detection orientation at any given moment. , They are respectively The high-precision data position relative to the skywave radar at any given time, including distance and azimuth. It is the root mean square. for High-precision data of the radial velocity relative to the skywave radar at all times.

8. The method according to claim 5, characterized in that, In step 4.4, the formula for calculating the decision parameter is as follows: in, l v To meet the requirements of this high-precision trajectory The number of waypoints , These represent the radial velocities at a set of adjacent moments for the detection track and the high-precision track, respectively. For detection batch number The number of waypoints corresponding to the longest high-precision track. This represents the number of waypoints for this high-precision track.

9. The method according to claim 1, characterized in that, Step 5 includes: Step 5.1: Obtain the total number of track points detected before matching. Total number of all matched detection tracks and the total number of corresponding waypoints. ; Step 5.2: Calculate the average radial velocity jitter for all matched data. Average single-track matching rate Overall matching rate .

10. The method according to claim 9, characterized in that, In step 5.2, The calculation formula is: in, This represents the number of trackpoints for a given matched probe track. This is the first of the exploration tracks. The difference between a radial velocity and the radial velocity at the nearest neighbor time point of the corresponding high-precision track. This is the average of all radial velocity differences along the probe track; The calculation formula is: in, For the detection batch number The number of track points matched by the detected track. To match the total number of track points of the previous probe track; The calculation formula is: 。

Images