Ship track prediction matching method based on radar track data

Abstract

The invention discloses a ship track prediction matching method based on radar track data. The method mainly comprises the following steps: carrying out preliminary cleaning on radar track point data to be processed by utilizing a longitude and latitude range; converting the latitude and longitude coordinates of the radar track points into projection coordinates of the latitude of the monitored sea area; selecting part of track points in front of the interruption point according to the time of the radar track points; further cleaning the radar track points by using the distance between the track points; selecting a plurality of track points before the interruption point from the remaining track points according to time; calculating the speed in the longitude direction and the speed in the latitude direction according to the navigational speed and the navigational direction of the track point; obtaining a time interval between the track points and a ship radar ID number; a plurality of predicted track points are obtained through Kalman filtering prediction; other ship track points closest to the predicted track point are searched, and radar ID numbers are recorded; searching a track which is determined to be interrupted by the radar ID number with the maximum appearance number; according to the method, ship radar track leakage point supplement and automatic connection can be realized.

Description

Technical Field

[0001] The present invention relates to the field of big data processing, and in particular to a ship trajectory prediction and matching method based on radar track data. Background Art

[0002] With the continuous development of shipping, maritime traffic is becoming increasingly busy. Improving the supervision of maritime navigation and identifying abnormal activities are of great significance for creating a good marine order and protecting the lives and property of the people.

[0003] With the development of information technology in regulatory units, radar, AIS, and Beidou data collection technologies have effectively supported regulatory operations, but some illegal ships often turn off AIS and satellite equipment to avoid detection. At this time, the radar system that monitors the sea area will become the only means to detect illegal ships. However, in actual applications, there are coverage gaps between radars, and the electromagnetic waves emitted by the radar will be blocked by islands and large ships. Sea conditions and weather will also affect the reception of radar signals. It is difficult for the system to continuously track ships in the sea area, causing interruptions in the ship's track.

[0004] In order to determine whether the track before and after the interruption belongs to the same ship, experienced law enforcement personnel are often required to make manual judgments, which consumes a lot of manpower and time and is prone to false detection and missed detection. This shows that relying solely on radar data for ship supervision still faces major challenges. Summary of the invention

[0005] In order to solve the problem that the ship track is often interrupted in the current process of ship supervision using radar ship data, and a large amount of manpower and time are required for manual judgment, the present invention provides a ship track prediction and matching method based on radar track data. The method uses a computer algorithm to intelligently process the received ship radar track data, uses the ship track before the interruption to predict, matches the prediction result with the track after the interruption, and then integrates the ship tracks before and after the interruption to achieve automatic connection of the ship radar track.

[0006] The purpose of the present invention can be achieved by adopting the following technical solutions:

[0007] A ship trajectory prediction and matching method based on radar track data, characterized in that it comprises the following steps:

[0008] S1. Preliminary cleaning of the radar track point data to be processed is performed, and according to the longitude and latitude range of the monitored sea area, the points whose longitude and latitude are out of range or located on land are removed;

[0009] S2, coordinate conversion is performed on the radar track points after preliminary cleaning, and the longitude and latitude coordinates are converted into the projection coordinates of the latitude of the monitored sea area;

[0010] S3, according to the time of the radar track point, select the M track points before the interruption point, and set them in order from the earliest to the last time as P1, P2, ..., P M The meridian coordinates of each point are y1, y2, ..., y M The latitudinal coordinates of each point are x1, x2, ..., x M ;

[0011] S4, further clean the radar track points, starting from i=1, for P i , P i+1 , P i+2 These three consecutive track points satisfy the following formula:

[0012]

[0013] Then determine the track point P i+1 It is an error point, so it is removed and then i+2 Start to repeat the above calculation, otherwise the track point P is determined i+1 is the normal point, and then i+1 Start repeating the above calculations;

[0014] S5. Select N track points before the interruption point from the track points remaining after further cleaning according to time, and set them in order from the earliest to the last time as P'1, P'2, ..., P' N The meridian coordinates of each point are y′1, y′2, ..., y′ N The latitudinal coordinates of each point are x′1, x′2, …, x′ N ;

[0015] S6. Obtain the speed and heading data of N track points, set them as v1, v2, ..., v N and θ1, θ2, ..., θ N , for the i-th track point P′ i , calculate its longitude direction speed and latitude speed As shown below:

[0016]

[0017]

[0018] S7. According to the selected track points, the time interval Δt between each point and the radar ID number of the ship are obtained. C ;

[0019] S8. Based on uniformly accelerated linear motion, the transfer matrix of Kalman filter is set as:

[0020]

[0021] Set the measurement matrix of the Kalman filter to:

[0022]

[0023] If the process noise coefficient of the Kalman filter is set to ε1, the corresponding Kalman filter process noise matrix is:

[0024]

[0025] If the Kalman filter measurement noise coefficient is set to ε2, the corresponding Kalman filter measurement noise matrix is:

[0026]

[0027] According to the existing track points and the above settings, Kalman filter is used to predict and obtain K predicted track points, which are recorded as Q1, Q2, ..., Q K ;

[0028] S9. For the i-th predicted track point Q i , find the nearest ship track point and record the radar ID number of the point, denoted as I i ;

[0029] S10, find I1, I2, ..., I K The radar ID number with the largest number is determined to be the ship track corresponding to the radar ID number after the interruption, and its radar ID number is reset to I C , thus completing the connection of the ship radar tracks.

[0030] As a preferred technical solution, in the step S2, the pyproj tool library is used to convert the longitude and latitude coordinates into projection coordinates of the latitude of the monitored sea area.

[0031] As a preferred technical solution, in step S8, the Kalman filter function of the opencv toolkit is used for prediction.

[0032] As a preferred technical solution, the number M of track points selected in step S3 should be no less than 4 times the number N of track points selected in step S5 to ensure that sufficient track points can still be selected after further cleaning.

[0033] As a preferred technical solution, the number K of track points predicted in step S8 should be no less than the number N of track points selected in step S5 to make full use of existing data.

[0034] Compared with the prior art, the present invention has the following advantages and effects:

[0035] 1. The present invention provides a ship trajectory prediction and matching method based on radar track data, which can realize the automatic connection of radar ship tracks, assist law enforcement personnel in completing the analysis of radar ship tracks, and effectively improve the efficiency of ship supervision work.

[0036] 2. The present invention performs preliminary cleaning of radar ship track points through longitude and latitude ranges, and further cleans radar ship track points through distances between points, thereby avoiding erroneous data caused by radar wave interference from being used for prediction, thereby improving the accuracy of track prediction.

[0037] 3. The present invention utilizes the Kalman filter algorithm to perform prediction based on the piecewise uniformly variable speed motion model, and the computational complexity is low, thereby effectively improving the application efficiency of the method. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is an overall flow chart of the radar ship track connection method based on orbit prediction and matching disclosed in the present invention;

[0039] Figure 2 It is an example diagram of predicted track points and actual ship radar tracks in an embodiment of the present invention. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0041] This embodiment discloses a ship trajectory prediction and matching method based on radar track data. The following takes the ship radar track data in Guangdong waters as an example as a preferred embodiment to introduce the implementation process of the present invention in detail. The overall implementation process is as follows: Figure 1 As shown, the specific implementation steps are as follows:

[0042] The radar track point data to be processed is preliminarily cleaned, and according to the longitude and latitude range of the monitored sea area, the points in the track points whose longitude and latitude are out of range or located on land are removed.

[0043] The radar track points after preliminary cleaning are converted into coordinates, and the pyproj tool library is used to convert the WGS 84 longitude and latitude coordinates into the WGS 84 / UTM zone 50N projection coordinates of the latitude of the monitored sea area.

[0044] According to the time of the radar track point, select M = 200 track points before the interruption point and set them in order of time from the earliest to the last as P1, P2, ..., P 200 The meridian coordinates of each point are y1, y2, ..., y 200 The latitudinal coordinates of each point are x1, x2, ..., x 200 .

[0045] Further cleaning of radar track points, starting from i=1, for P i , P i+1 , P i+2 These three consecutive track points satisfy the following formula:

[0046]

[0047] Then determine the track point P i+1 It is an error point, so it is removed and then i+2 Start to repeat the above calculation, otherwise the track point P is determined i+1 is the normal point, and then i+1 The above calculation is repeated. The above operation is repeated to remove 38 wrong track points.

[0048] From the remaining track points after further cleaning, select N = 50 track points before the interruption point according to time, and set them in order from the earliest to the last time as P'1, P'2, ..., P' 50 The meridian coordinates of each point are y′1, y′2, ..., y′ 50 The latitudinal coordinates of each point are x′1, x′2, …, x′ 50 .

[0049] Get the speed and heading data of 50 track points, set as v1, v2, ..., v 50 and θ1, θ2, ..., θ 50 , for the i-th track point P′1, calculate its longitude direction speed and latitude speed As shown below:

[0050]

[0051]

[0052] According to the selected track points, the time interval between each point Δt = 8 seconds and the ship's radar ID number I C =001.

[0053] Based on uniformly accelerated linear motion, the transfer matrix of Kalman filter is set as:

[0054]

[0055] Set the measurement matrix of the Kalman filter to:

[0056]

[0057] The process noise coefficient of the Kalman filter is set to ε1 = 0.03, and the corresponding Kalman filter process noise matrix is:

[0058]

[0059] The measurement noise coefficient of the Kalman filter is set to ε2 = 0.01, and the corresponding Kalman filter measurement noise matrix is:

[0060]

[0061] According to the existing track points and the above settings, Kalman filter is used to predict and obtain K = 50 predicted track points, which are recorded as Q1, Q2, ..., Q 50 .

[0062] For the i-th predicted track point Q i , find the nearest ship track point and record the radar ID number of the point, denoted as I i ;

[0063] Find I1, I2, ..., I 50 The radar ID number with the largest number is 003, so the ship track corresponding to the radar ID number is determined to be the track after the interruption, and the radar ID number is reset from 003 to 001, thus completing the connection of the ship radar track. The predicted track point and the track after the interruption point are as follows Figure 2 As shown, the dots are the predicted ship track points and the solid line is the ship radar track after the interruption point.

[0064] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.

Claims

1. A ship trajectory prediction and matching method based on radar track data, characterized in that: The following steps are involved: S1. Perform preliminary cleaning on the radar track point data to be processed. According to the longitude and latitude range of the monitored sea area, remove the track points whose longitude and latitude are out of range or located on land; S2. Convert the coordinates of the radar track points after preliminary cleaning to convert the longitude and latitude coordinates into the projection coordinates of the latitude of the monitored sea area; S3, according to the time of the radar track point, select the M track points before the interruption point, and set them in order of time from the earliest to the last as P1, P2, ..., P M The longitudinal coordinates of each point are y1, y2, ..., y M The latitudinal coordinates of each point are x1, x2, ..., x M ; S4, further clean the radar track points, starting from i=1, for P i 、P i+1 、P i+2 These three consecutive track points satisfy the following formula: Then determine the track point P i+1 For error points, remove them and then i+2 Start to repeat the above calculation, otherwise the track point P is determined i+1 is the normal point, and then i+1 Start repeating the above calculations; S5, select N track points before the interruption point from the track points remaining after further cleaning according to time, and set them in order from the earliest to the last time as P1', P2', ..., P N ′, the longitudinal coordinates of each point are y1′, y′2, ..., y′ N The latitudinal coordinates of each point are x1′, x′2, …, x′ N ; S6. Obtain the speed and heading data of N track points, set as v1, v2, ..., v N and θ1, θ2, ..., θ N , for the i-th track point P i ′, calculate its longitude speed and latitudinal speed As shown in the following formula: S7. Based on the selected track points, obtain the time interval Δt between each point and the radar ID number of the ship C ; S8. Based on uniformly accelerated linear motion, the transfer matrix of the Kalman filter is set to: Set the measurement matrix of the Kalman filter to: If the process noise coefficient of the Kalman filter is set to ε1, the corresponding Kalman filter process noise matrix is: If the Kalman filter measurement noise coefficient is set to ε2, the corresponding Kalman filter measurement noise matrix is: According to the existing track points and the above settings, Kalman filter is used to predict and obtain K predicted track points, which are recorded as Q1, Q2, ..., Q K ; S9. For the i-th predicted track point Q i , find the nearest ship track point and record the radar ID number of the point, which is recorded as I i ; S10, search for I1, I2, ..., I K The radar ID number with the largest number is determined to be the ship track corresponding to the radar ID number after the interruption, and its radar ID number is reset to I C , thus completing the connection of the ship radar tracks.

2. The ship trajectory prediction and matching method based on radar track data according to claim 1, characterized in that: In the step S2, the pyproj tool library is used to convert the longitude and latitude coordinates into the projection coordinates of the latitude of the monitored sea area.

3. The ship trajectory prediction and matching method based on radar track data according to claim 1, characterized in that: In step S8, the Kalman filter function of the opencv toolkit is used for prediction.

4. The ship trajectory prediction and matching method based on radar track data according to claim 1, characterized in that: The number M of track points selected in step S3 should be no less than four times the number N of track points selected in step S5 to ensure that enough track points can still be selected after further cleaning.

5. The ship trajectory prediction and matching method based on radar track data according to claim 1, characterized in that: The number K of track points predicted in step S8 should be no less than the number N of track points selected in step S5 to fully utilize the existing data.

Images