A radar and AIS data fusion method based on multi-factor euclidean distance correlation

Abstract

The application discloses a radar and AIS data fusion method based on a multi-factor Euclidean distance correlation, which comprises the following steps: 1, radar data preprocessing is realized through data processing methods such as outlier elimination, time unification and space unification; 2, the target offset of the radar data is calculated based on the AIS target data, the data adaptive correction parameter and the checking method are established, and the radar data correction is realized; 3, the corrected radar data and the AIS data are subjected to track correlation, and the distance between the latest target data of the two sensors is taken as the judgment basis to perform target coarse correlation; 4, the track fine correlation is performed by using a multi-factor Euclidean distance correlation algorithm, and finally the data fusion is completed. The application can realize the multi-source data processing, correction and fusion of the marine target, has certain interpretability, and provides a reliable data processing method for the real-time tracking, supervision and track backtracking of the marine ship target.

Description

Technical Field

[0001] This invention relates to the technical field of radar and AIS data fusion, and in particular to a radar and AIS data fusion method based on multi-factor Euclidean range correlation. Background Technology

[0002] Maritime radar and AIS systems provide crucial navigation data for ships. During radar data acquisition, external conditions such as transmission channel congestion, obstruction, and sea clutter interference can inevitably lead to issues like lost radar targets and temporary track gaps. AIS data is rich, including call sign, ship position, speed, heading, time, length, and other ship-related information, with no blind spots within its reception range. While ship position, speed, and time in AIS data are primarily based on the Global Positioning System (GPS) and have high accuracy, call sign and length information are fixed. However, the data transmission period of AIS terminals is not fixed, resulting in periods of data stagnation. Therefore, fusing radar and AIS data can significantly improve the quality of ship navigation data.

[0003] Common radar and AIS data fusion methods have two main problems that need to be optimized:

[0004] 1. The preprocessing of radar and AIS data was incomplete, and data offset was not corrected.

[0005] 2. When radar and AIS tracks are correlated, multiple factors need to be considered, and the weights of these multiple factors need to be allocated in a more scientific way.

[0006] Therefore, research should focus on preprocessing techniques for raw radar and AIS data, new methods for track association, and weight allocation methods for various track factors to address the existing problems in current radar and AIS data fusion technologies. Summary of the Invention

[0007] The technical problem to be solved by the present invention is to address the shortcomings of the prior art by providing a radar and AIS data fusion method based on multi-factor Euclidean distance correlation. This radar and AIS data fusion method based on multi-factor Euclidean distance correlation can realize the scientific preprocessing of radar data and AIS data, establish an effective data correction model and track correlation fusion model, and provide reliable data support for real-time tracking, monitoring and track tracing of maritime targets.

[0008] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0009] A radar and AIS data fusion method based on multi-factor Euclidean distance correlation includes the following steps.

[0010] Step 1, Spatiotemporal Unification: The radar data and AIS data monitored in the same sea area will be unified in time and space; among them, the space will be unified to the Cartesian coordinate system.

[0011] Step 2, Radar Data Correction: Based on AIS data, the radar data is corrected to prevent target deviation from the radar data.

[0012] Step 3, Coarse Track Correlation: The distance between the corrected radar data and AIS data is used as the basis for coarse correlation of the target track.

[0013] Step 4, Fine-grained Track Association: Using a multi-factor Euclidean distance association algorithm, the target tracks after coarse association are fine-grained, thereby completing the fusion of radar data and AIS data; the multi-factor Euclidean distance association algorithm uses the combination of Euclidean distances of multiple factors related to ship navigation dynamics as the basis for judging fine association.

[0014] Step 3, the method for coarse correlation of tracks, includes the following steps:

[0015] Step 3-1: Set the distance threshold.

[0016] Step 3-2, Coarse Association: When the distance between the corrected radar data and AIS data is less than the distance threshold set in Step 3-1, it is determined that the current two target tracks satisfy coarse association; at this time, the AIS data is coarsely associated and matched to the radar data to form a coarsely associated track dataset.

[0017] Step 4, the method for fine-grained track correlation, includes the following steps:

[0018] Step 4-1: Select navigation dynamics influencing factors: Select n influencing factors related to the ship's navigation dynamics; where n≥2.

[0019] Step 4-2: Calculate the one-factor Euclidean distance: Let dis i Dis represents the Euclidean distance between the latest data point in the coarsely correlated trajectory and the i-th influencing factor of the target point to be correlated, where 1≤i≤n; i The calculation formula is:

[0020]

[0021] In the formula, m represents the total dimension of the data for the i-th influencing factor; j represents the j-th dimension of the data for the i-th influencing factor.

[0022] x j This represents the j-th dimension of the latest data point in the coarse correlation track, which is the i-th influencing factor.

[0023] y jThis represents the j-th dimension of the target point to be associated at the latest moment in the i-th influencing factor.

[0024] Step 4-3: Calculate the multi-factor Euclidean distance dis. The specific calculation formula is as follows:

[0025] dis = W1 × dis1 + W2 × dis2 + ... + W i ×dis i +…+W n ×dis n

[0026] in:

[0027] W1+W2+…W i +…+W n =1

[0028] In the formula, W1, W2, W i and W n These are the weighting coefficients of n influencing factors.

[0029] Step 4-4, Fine Association: Compare the multi-factor Euclidean distance dis calculated in Step 4-3 with the set multi-factor Euclidean distance threshold dis0; when dis < dis0, it is determined that the current two target tracks satisfy fine association; at this time, the target points to be associated are matched to the coarse association tracks to form a fused fine association track dataset.

[0030] In step 4-3, the weighting coefficients W1, W2, and W... i and W n The solution was obtained using the Analytic Hierarchy Process (AHP).

[0031] In step 4, n = 4, and the four factors are the distance between the two targets D, the ground speed SOG, the ground heading COG, and the true heading Hdg.

[0032] Step 4 involves using the Analytic Hierarchy Process (AHP) to solve for the weight coefficients W1, W2, W3, and W4, and includes the following steps:

[0033] Step 4-3a: Construct the judgment matrix P: Using the four factors U(D), U(SOG), U(COG), and U(Hdg) as the judgment matrix criteria, construct the judgment matrix P. The specific expression is as follows:

[0034]

[0035] And let:

[0036] u1 = U(D)

[0037] u2 = U(SOG)

[0038] u3 = U(COG)

[0039] u4=U(Hdg)

[0040] In the formula, u 11 This indicates the relative importance of the distance U(D) between two targets to themselves.

[0041] u 14 This indicates the importance of the distance U(D) between two targets relative to the true heading U(Hdg).

[0042] u 41 This indicates the importance of the true heading U(Hdg) relative to the distance U(D) between the two targets.

[0043] u 44 This indicates the relative importance of the true bow direction U(Hdg) to itself.

[0044] Step 4-3b: Calculate the weight coefficients: Calculate the largest eigenvalue λ of the judgment matrix P. max The eigenvector w is obtained by normalizing w, and then the normalized value W is obtained.

[0045] Step 4-3c, Consistency check: The maximum eigenvalue λ calculated in step 4-3b is used for consistency verification. max Perform a consistency check. If the judgment matrix P passes the consistency check, then the normalized value W in step 4-3b is the weight coefficients W1, W2, W3 and W4 obtained by solving.

[0046] Step 2, the radar data correction method includes the following steps:

[0047] Step 2-1: Divide the radar detection range into q sector units, where q ≥ 8.

[0048] Step 2-2: Set the initial distance correction coefficient for each sector: Based on the AIS data and the position of each sector, set an initial distance correction coefficient for each sector.

[0049] Steps 2-3: First correction: The radar position data of each sector is corrected using the corresponding initial range correction coefficient to obtain the radar position data after the first correction.

[0050] Step 2-4, First Error Calculation: Compare the radar position data after the first correction with the AIS data and perform a first error calculation; if the first error value does not exceed the set first error threshold, the radar data correction is complete; otherwise, proceed to step 2-5.

[0051] Steps 2-5: Correct the distance correction coefficient, including the following steps:

[0052] Step 2-5a, Verification: Use the cosine function to perform distance verification on the radar position data and AIS data after one correction that exceed the error threshold.

[0053] Step 2-5b, Correlation Judgment: If the distance verification in step 2-5a is normal, the two targets are judged to be correlated and proceed to step 2-5c; otherwise, the two targets are judged to be uncorrelated.

[0054] Step 2-5c: Correcting the range correction coefficient: By comparing the radar position data before and after the first correction, the initial range correction coefficient is corrected and updated; let the corrected range correction coefficient be paranew; paranew includes the new range correction coefficient x component paranew x And the new distance correction coefficient y component paranew y Then paranew x and paranew y The corrected formulas are as follows:

[0055]

[0056]

[0057] In the formula, A c x and Ax represent the radar longitude after one correction and the radar longitude after one correction, respectively.

[0058] A c y and Ay represent the radar latitude after one correction and the radar latitude after one correction, respectively.

[0059] paraold x and paraold y These represent the x-component and y-component of the initial distance correction coefficient, respectively.

[0060] In step 2-1, the radar detection range is divided into 72 sector units, with each sector angle divided into a unit of 15° and each distance unit divided into a range unit of 8 nautical miles.

[0061] In step 1, before the spatiotemporal unification begins, outlier removal is performed on both radar data and AIS data. The outlier removal method is as follows: the distance L between the latest target location data and the historical target location data in the radar data or AIS data is calculated, and the target data location is estimated. When the distance L is greater than the pre-set outlier judgment threshold, the latest target location data is judged to be an outlier and is removed.

[0062] In step 1, when time is unified, the linear interpolation method is used to perform interpolation calculations to fill in the missing data segments due to the inconsistent update frequencies of AIS data and radar data; when space is unified, the Gauss-Kruger projection algorithm is used to convert AIS data from geographic coordinates to Cartesian coordinates, and trigonometric functions are used to convert radar data from polar coordinates to Cartesian coordinates.

[0063] The present invention has the following beneficial effects:

[0064] This invention is based on data preprocessing methods such as outlier removal, linear interpolation, and Gauss-Kruger projection. It constructs a mathematical model of a multi-factor Euclidean distance correlation fusion algorithm based on data correction and coarse correlation processing. The reliability, accuracy, and advancement of the data preprocessing and fusion algorithm are verified through data comparison. This invention provides a reference data processing method, data support, and system for shipbuilding and maritime-related departments. Attached Figure Description

[0065] Figure 1 The flowchart of a radar and AIS data fusion method based on multi-factor Euclidean distance correlation according to the present invention is shown.

[0066] Figure 2 This diagram illustrates the radar and AIS track association model based on multi-factor Euclidean distance in this invention.

[0067] Figure 3 This is a comparison of the effects of single-factor Euclidean distance correlation and multi-factor Euclidean distance correlation at times T1 and T2 in the trajectory correlation process of this invention. Among them, (a) is the effect of single-factor Euclidean distance correlation at time T1; (b) is the effect of multi-factor Euclidean distance correlation at time T1; (c) is the effect of single-factor Euclidean distance correlation at time T2; and (d) is the effect of multi-factor Euclidean distance correlation at time T2. Detailed Implementation

[0068] The present invention will now be described in further detail with reference to the accompanying drawings and specific preferred embodiments.

[0069] In the description of this invention, it should be understood that the terms "left side," "right side," "upper part," "lower part," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. "First," "second," etc., do not indicate the importance of the components, and therefore should not be construed as a limitation of this invention. The specific dimensions used in this embodiment are only for illustrating the technical solution and do not limit the scope of protection of this invention.

[0070] like Figure 1As shown, a radar and AIS data fusion method based on multi-factor Euclidean distance correlation includes the following steps.

[0071] Step 1: Spatiotemporal unification

[0072] Step 1-1: Outlier Removal

[0073] Raw AIS and radar data inevitably contain outliers. Traditional fusion algorithms haven't processed all the data, leading to occasional large fluctuations in data accuracy. Outlier removal during data preprocessing can improve data accuracy.

[0074] The preferred outlier removal method is as follows: calculate the distance L between the latest target location data and the historical target location data in radar data or AIS data, and estimate the target data location; when the distance L is greater than the preset outlier judgment threshold, the latest target location data is judged to be an outlier and removed.

[0075] Step 1-2, Time Consistency: By using linear interpolation to address the inconsistency in update frequencies between AIS and radar data, missing data segments are filled in to achieve consistent update frequencies and ensure time consistency.

[0076] Taking location information as an example, in t k and t k+1 The position information at time points are A(Lon1,Lat1) and B(Lon2,Lat2), respectively. When t∈(t k ,t k+1 )hour, t The time location information (Lon,Lat) can be expressed as Equation (1).

[0077]

[0078] Steps 1-3, Spatial Unification: The Gauss-Kruger projection algorithm is used to convert AIS data from geographic coordinates to Cartesian coordinates, and trigonometric functions are used to convert radar data from polar coordinates to Cartesian coordinates.

[0079] Step 2, Radar Data Correction

[0080] Due to the radar's inherent calibration settings, the raw radar data will exhibit a zero-bias error. This is a fixed error caused by the hardware system, and traditional methods typically address it through the calibration function of the data processing subsystem, requiring manual correction based on experience. This step constructs an adaptive partitioned correction parameter and a calculation iterative verification method. Using AIS data as a benchmark, it adaptively corrects the radar data to prevent target offset from occurring in the radar data.

[0081] The radar data correction method described above preferably includes the following steps.

[0082] Step 2-1: Sector Division: Divide the radar detection range into q sector units; where q ≥ 8. In this embodiment, the radar detection range is divided into units based on sector angle, with each unit being 15° and each unit being a range unit being 8 nautical miles (taking a certain type of solid-state radar as an example, with the detection mode range selected as 24 nautical miles), for a total of 72 sector units.

[0083] Step 2-2: Set the initial distance correction coefficient for each sector: Based on AIS data and the location of each sector, set an initial distance correction coefficient for each sector. The initial distance correction coefficient includes an x-component and a y-component.

[0084] Steps 2-3: First correction: The radar position data of each sector is corrected using the corresponding initial range correction coefficient to obtain the radar position data after the first correction.

[0085] Step 2-4, First Error Calculation: Compare the radar position data after the first correction with the AIS data and perform a first error calculation; if the first error value does not exceed the set first error threshold, the radar data correction is complete; otherwise, proceed to step 2-5.

[0086] Steps 2-5: Correct the distance correction coefficient, including the following steps:

[0087] Step 2-5a, Verification: Use the cosine function to perform distance verification on the radar position data and AIS data after one correction that exceed the error threshold.

[0088] Step 2-5b, Correlation Judgment: If the distance verification in step 2-5a is normal, the two targets are judged to be correlated and proceed to step 2-5c; otherwise, the two targets are judged to be uncorrelated.

[0089] Step 2-5c: Correcting the range correction coefficient: By comparing the radar position data before and after the first correction, the initial range correction coefficient is corrected and updated; let the corrected range correction coefficient be paranew; paranew includes the new range correction coefficient x component paranew x And the new distance correction coefficient y component paranew y Then paranew x and paranew y The corrected formulas are as follows:

[0090]

[0091]

[0092] In the formula, A c x and Ax represent the radar longitude after one correction and the radar longitude after one correction, respectively.

[0093] A c y and Ay represent the radar latitude after one correction and the radar latitude after one correction, respectively.

[0094] paraold x and paraold y These represent the x-component and y-component of the initial distance correction coefficient, respectively.

[0095] Step 3, Coarse Track Correlation: The distance between the corrected radar data and AIS data is used as the basis for coarse correlation of the target track.

[0096] The above-mentioned method for coarse correlation of tracks preferably includes the following steps.

[0097] Step 3-1: Set the distance threshold.

[0098] Step 3-2, Coarse Association: When the distance between the corrected radar data and AIS data is less than the distance threshold set in Step 3-1, it is determined that the current two target tracks satisfy coarse association; at this time, the AIS data is coarsely associated and matched to the radar data to form a coarsely associated track dataset.

[0099] Step 4, Fine-grained Track Association: Using a multi-factor Euclidean distance association algorithm, the target tracks after coarse association are fine-grained, thereby completing the fusion of radar data and AIS data; the multi-factor Euclidean distance association algorithm uses the combination of Euclidean distances of multiple factors related to ship navigation dynamics as the basis for judging fine association.

[0100] The above-mentioned method for fine correlation of flight paths preferably includes the following steps.

[0101] Step 4-1: Select navigation dynamics influencing factors: Select n influencing factors related to the ship's navigation dynamics; where n≥2. In this embodiment, n=4 is preferred, and the four factors are the distance between the two targets D, the speed SOG over the ground, the heading COG over the ground, and the true heading Hdg.

[0102] Step 4-2: Calculate the one-factor Euclidean distance: Let dis i Dis represents the Euclidean distance between the latest data point in the coarsely correlated trajectory and the i-th influencing factor of the target point to be correlated, where 1≤i≤n; i The calculation formula is:

[0103]

[0104] In the formula, m represents the total dimension of the data for the i-th influencing factor; j represents the j-th dimension of the data for the i-th influencing factor.

[0105] x j This represents the j-th dimension of the latest data point in the coarse correlation track, which is the i-th influencing factor.

[0106] y j This represents the j-th dimension of the target point to be associated at the latest moment in the i-th influencing factor.

[0107] Step 4-3: Calculate the multi-factor Euclidean distance dis. The specific calculation formula is as follows:

[0108] dis = W1 × dis1 + W2 × dis2 + ... + W i ×dis i +…+W n ×dis n

[0109] in:

[0110] W1+W2+…W i +…+W n =1

[0111] In the formula, W1, W2, W i and W n These are the weighting coefficients of n influencing factors.

[0112] The aforementioned weighting coefficients W1, W2, W i and W n The preferred method is the Analytic Hierarchy Process (AHP) to obtain the solution.

[0113] In this embodiment, since n=4, the method of solving the weight coefficients W1, W2, W3 and W4 using the analytic hierarchy process (AHP) preferably includes the following steps.

[0114] Step 4-3a: Construct the judgment matrix P: Using the four factors U(D), U(SOG), U(COG), and U(Hdg) as the judgment matrix criteria, construct the judgment matrix P. The specific expression is as follows:

[0115]

[0116] And let:

[0117] u1 = U(D)

[0118] u2 = U(SOG)

[0119] u3 = U(COG)

[0120] u4=U(Hdg)

[0121] In the formula, u 11 This indicates the relative importance of the distance U(D) between two targets to themselves.

[0122] u 14 This indicates the importance of the distance U(D) between two targets relative to the true heading U(Hdg).

[0123] u 41 This indicates the importance of the true heading U(Hdg) relative to the distance U(D) between the two targets.

[0124] u 44 This indicates the relative importance of the true bow direction U(Hdg) to itself.

[0125] Let u be any factor in the judgment matrix P. kl The values ​​of are taken according to Table 1 below, where 1≤k≤4 and 1≤l≤4.

[0126] Table 1 Definition of Matrix Scale

[0127]

[0128] In this embodiment, a detailed calculation of the weighting process for distance D, ground speed SOG, ground heading COG, and true heading Hdg is performed using data examples. The data used is only for explaining the calculation method, including but not limited to the calculation parameters used below.

[0129] Based on Table 1, the values ​​of the four factors are obtained, as shown in Table 2 below.

[0130] Table 2

[0131]

[0132] Therefore, the constructed judgment matrix is ​​shown in the following equation:

[0133]

[0134] Step 4-3b: Calculate the weight coefficients: Calculate the largest eigenvalue λ of the judgment matrix P. max The eigenvector w is obtained by normalizing w, and then the normalized value W is obtained.

[0135] In this embodiment, the solution is:

[0136] λ max =4.1113

[0137] w=[0.7455 0.5760 0.2212 0.2520] T

[0138] Normalizing the eigenvectors yields:

[0139] W = [W1 W2 W3 W4] T =[0.4154 0.3209 0.1233 0.1404] T

[0140] Step 4-3c, Consistency check: The maximum eigenvalue λ calculated in step 4-3b is used for consistency verification. max Perform a consistency check. If the judgment matrix P passes the consistency check, then the normalized value W in step 4-3b is the weight coefficients W1, W2, W3 and W4 obtained by solving.

[0141] In this embodiment, the judgment matrix P passes the consistency test, so the weights calculated above satisfy the random consistency ratio condition and can be used as the weight allocation result.

[0142] Step 4-4, Fine Association: Compare the multi-factor Euclidean distance dis calculated in Step 4-3 with the set multi-factor Euclidean distance threshold dis0; when dis < dis0, it is determined that the current two target tracks satisfy fine association; at this time, the target points to be associated are matched to the coarse association tracks to form a fused fine association track dataset.

[0143] The preferred embodiments of the present invention have been described in detail above. However, the present invention is not limited to the specific details in the above embodiments. Within the scope of the technical concept of the present invention, various equivalent transformations can be made to the technical solutions of the present invention, and these equivalent transformations all fall within the protection scope of the present invention.

Claims

1. A radar and AIS data fusion method based on multi-factor Euclidean range correlation, characterized in that: Includes the following steps: Step 1, Spatiotemporal Unification: The radar data and AIS data monitored in the same sea area will be unified in terms of time and space; among them, the space will be unified to the Cartesian coordinate system. Step 2, Radar Data Correction: Using AIS data as a reference, the radar data is corrected to prevent target deviation. The radar data correction method includes the following steps: Step 2-1: Divide the radar detection range into q sector units; where q ≥ 8; Step 2-2: Set the initial distance correction coefficient for each sector: Based on the AIS data and the position of each sector, set an initial distance correction coefficient for each sector. Steps 2-3: First correction: The radar position data of each sector is corrected using the corresponding initial range correction coefficient to obtain the radar position data after the first correction. Step 2-4, First Error Calculation: Compare the radar position data after the first correction with the AIS data and perform a first error calculation; if the first error value does not exceed the set first error threshold, the radar data correction is complete; otherwise, proceed to step 2-5. Steps 2-5: Correct the distance correction coefficient, including the following steps: Step 2-5a, Verification: Use the cosine function to verify the distance of the radar position data and AIS data after one correction that exceed the first error threshold; Step 2-5b, Correlation Judgment: If the distance verification in Step 2-5a is normal, the two targets are judged to be correlated, and proceed to Step 2-5c; otherwise, the two targets are judged to be uncorrelated. Step 2-5c: Correct the range correction coefficient: By comparing the radar position data before and after the first correction, the initial range correction coefficient is corrected and updated; let the corrected range correction coefficient be... ; Including the new distance correction factor x component And the new distance correction coefficient y component ;but and The corrected formulas are as follows: ； ； In the formula, , This indicates the radar longitude after one correction and the radar longitude after one correction. , This indicates the radar latitude after one correction and the radar latitude after one correction. and These represent the x-component and y-component of the initial distance correction coefficient, respectively. Step 3, Coarse Track Correlation: The distance between the corrected radar data and AIS data is used as the basis for coarse correlation of the target track; Step 4, Fine-grained Track Association: Using a multi-factor Euclidean distance association algorithm, the target tracks after coarse association are fine-grained, thereby completing the fusion of radar data and AIS data; the multi-factor Euclidean distance association algorithm uses the combination of Euclidean distances of multiple factors related to ship navigation dynamics as the basis for judging fine association.

2. The radar and AIS data fusion method based on multi-factor Euclidean range correlation according to claim 1, characterized in that: Step 3, the method for coarse correlation of tracks, includes the following steps: Step 3-1: Set the distance threshold; Step 3-2, Coarse Association: When the distance between the corrected radar data and AIS data is less than the distance threshold set in Step 3-1, it is determined that the current two target tracks satisfy coarse association; at this time, the AIS data is coarsely associated and matched to the radar data to form a coarsely associated track dataset.

3. The radar and AIS data fusion method based on multi-factor Euclidean range correlation according to claim 2, characterized in that: Step 4, the method for fine-grained track correlation, includes the following steps: Step 4-1: Select factors influencing navigation dynamics: Select n factors related to ship navigation dynamics; where n≥2; Step 4-2, Calculate the single-factor Euclidean distance: Let... This indicates the latest data point in the coarsely correlated track and the target point to be correlated. The Euclidean distance between the influencing factors, where 1 ≤ ≤ ;but The calculation formula is: ； In the formula, Indicates the first Total dimensions of data for each influencing factor; Indicates the first The first influencing factor Dimensional data; This indicates that the latest data point in the coarse correlation track is at the [number]th [time]. The first of the influencing factors Dimensional data; This indicates that the target point to be associated at the latest moment is at the th... The first of the influencing factors Dimensional data; Step 4-3: Calculate the multi-factor Euclidean distance The specific calculation formula is as follows: ； in: ； In the formula, These are the weighting coefficients of the n influencing factors; Step 4-4, Fine-grained correlation: Calculate the multi-factor Euclidean distance in step 4-3. The set multi-factor Euclidean distance threshold Compare and judge; when < If the two target tracks satisfy fine correlation, then the target points to be correlated are matched to the coarse correlation tracks to form a fused fine correlation track dataset.

4. The radar and AIS data fusion method based on multi-factor Euclidean range correlation according to claim 3, characterized in that: In step 4-3, the weighting coefficients The solution was obtained using the Analytic Hierarchy Process (AHP).

5. The radar and AIS data fusion method based on multi-factor Euclidean range correlation according to claim 4, characterized in that: In step 4, n=4, and the four factors are the distance between the two targets D, the ground speed SOG, the ground heading COG, and the true heading Hdg.

6. The radar and AIS data fusion method based on multi-factor Euclidean range correlation according to claim 5, characterized in that: In step 4, the weight coefficients are solved using the Analytic Hierarchy Process (AHP). The method includes the following steps: Step 4-3a, Construct the judgment matrix P: with , , and The four factors form the judgment matrix criterion. A judgment matrix P is constructed, with the following specific expression: ； And let: ； ； ； ； In the formula, Indicates the distance between two targets Its relative importance to itself; Indicates the distance between two targets relative to true bow The degree of importance; Indicates true bow direction Relative distance between two targets The degree of importance; Indicates true bow direction Its relative importance to itself; Step 4-3b: Calculate the weight coefficients: Calculate the largest eigenvalue of the judgment matrix P. eigenvectors and to After normalization, we obtain the normalized value. ; Step 4-3c, Consistency check: The largest eigenvalue calculated in step 4-3b is used for... Perform a consistency check. If the judgment matrix P passes the consistency check, then the normalized value in step 4-3b is... That is, the weight coefficients obtained by solving. .

7. The radar and AIS data fusion method based on multi-factor Euclidean range correlation according to claim 1, characterized in that: In step 2-1, the radar detection range is divided into 72 sector units, with each sector angle divided into a unit of 15° and each distance unit divided into a range unit of 8 nautical miles.

8. The radar and AIS data fusion method based on multi-factor Euclidean range correlation according to claim 1, characterized in that: In step 1, before the spatiotemporal unification begins, outlier removal is performed on both radar data and AIS data. The outlier removal method is as follows: the distance L between the latest target location data and the historical target location data in the radar data or AIS data is calculated, and the target data location is estimated. When the distance L is greater than the pre-set outlier judgment threshold, the latest target location data is judged to be an outlier and is removed.

9. The radar and AIS data fusion method based on multi-factor Euclidean range correlation according to claim 1 or 8, characterized in that: In step 1, when time is unified, the linear interpolation method is used to perform interpolation calculations to fill in the missing data segments due to the inconsistent update frequencies of AIS data and radar data; when space is unified, the Gauss-Kruger projection algorithm is used to convert AIS data from geographic coordinates to Cartesian coordinates, and trigonometric functions are used to convert radar data from polar coordinates to Cartesian coordinates.

Images