Ship movement data-based shipping traffic network extraction method and device, and storage medium

Abstract

The invention relates to a shipping traffic network extraction method and device based on ship movement data and a storage medium, and the method comprises the steps: S1, extracting data points of which the navigation states are anchoring and mooring in AIS data as basic ship berthing point data, and obtaining a plurality of berthing regions; s2, generating a plurality of effective tracks; s3, for each effective track, obtaining a plurality of staying areas, and fusing the parking areas and the staying areas to obtain a node set of the shipping traffic network; s4, based on the effective tracks, cyclically judging whether navigation tracks exist among the nodes or not, if one navigation track exists, taking the ship track as an arc between the two nodes, and if multiple tracks exist, performing track fusion by adopting a DTW algorithm to obtain a unique arc among the nodes; and S5, constructing the shipping traffic network based on the obtained node set and the arcs among the nodes. Compared with the prior art, the method has the advantages of high calculation speed, high accuracy and the like.

Description

Technical Field

[0001] The present invention relates to the field of maritime ship shipping management, and in particular to a shipping traffic network extraction method, device and storage medium based on ship movement data. Background Art

[0002] In pursuit of optimal fuel consumption, ships do not frequently change direction and try to pursue the shortest route, which means that there is a universal shipping traffic pattern, namely the shipping traffic network.

[0003] The extraction of shipping traffic networks helps to realize shipping traffic status analysis, shipping network flexibility analysis, port accessibility analysis and ship route search, providing support for corporate decision-making.

[0004] The shipping traffic network relies on ship movement data. Since 2004, the International Maritime Organization (IMO) has stipulated that cargo ships with a gross tonnage of more than 300 and all passenger ships on international voyages must install the Automatic Identification System (AIS). The AIS system provides a data foundation for ship behavior analysis and maritime management, including collision avoidance research, anomaly detection, behavior analysis, knowledge extraction, tracking and prediction, and also provides assistance for the mining of traffic networks.

[0005] For example, Chinese patent CN117785963A provides a method and system for mining a maritime traffic network based on open source AIS data, the method comprising: preprocessing AIS data; using geographic information and motion information to identify and classify docking trajectories, extracting ship docking areas containing semantic distribution of docking functions based on docking trajectories; by screening the trajectory points at the ship turning point, spatially dividing the area of ​​interest, and using genetic algorithms to extract the ship turning area; the collection of the ship docking area and the ship turning area is used as a feature area; constructing a directed graph based on the feature area, removing noise data, and generating route boundaries and characterizing routes through kernel density estimation and triangulation methods. The advantages of the present invention are: high extraction accuracy, low manual supervision cost, and rich regular semantics; while greatly accelerating the mining efficiency of the turning area, it can ensure that the mining accuracy will not be affected by the data density distribution.

[0006] However, compared with the present application, the steps of this method are more complicated and the method is more complex, but the difference in results is not obvious. Summary of the invention

[0007] The purpose of the present invention is to provide a method, device and storage medium for extracting a shipping traffic network based on ship movement data.

[0008] The purpose of the present invention can be achieved by the following technical solutions:

[0009] A shipping traffic network extraction method based on ship movement data, comprising:

[0010] Step S1: extracting data points with navigation status of anchoring and mooring in AIS data as basic ship berthing point data, and performing clustering processing using density-based clustering to obtain multiple berthing areas;

[0011] Step S2: According to the AIS data points, the track is divided based on the ship code, the navigation destination and the time interval of adjacent data points to generate multiple valid tracks;

[0012] Step S3: for each valid trajectory, select the data point whose navigation status is "navigating", select multiple waypoints based on ground speed, heading difference and speed difference, and cluster the waypoints obtained based on all valid trajectories using a density-based clustering method to obtain multiple stop areas, and merge the parking areas and stop areas to obtain a node set of the shipping traffic network;

[0013] Step S4: Based on the valid trajectory, it is cyclically determined whether there is a navigation trajectory between the nodes. If there is a navigation trajectory, the ship trajectory is used as the arc between the two nodes. If there are multiple trajectories, the DTW algorithm is used to fuse the trajectories to obtain a unique arc between the nodes.

[0014] Step S5: construct a shipping transportation network based on the obtained node set and arcs between nodes.

[0015] The step S1 comprises:

[0016] Step S1-1: Acquire AIS data and select data points with navigation status of anchoring and mooring as basic ship berthing point data;

[0017] Step S1-2: performing clustering processing using a density-based clustering method to obtain a plurality of first clusters;

[0018] Step S1-3: taking the center of each first cluster as the center of the first cluster, and taking the distance from the point with the largest distance from the center in the first cluster to the center as the radius of the first cluster, and constructing a parking area based on the center coordinates and radius of the first cluster;

[0019] Step S1-4: All parking areas are grouped into a parking area set.

[0020] The density-based clustering method is the DBSCAN algorithm.

[0021] In the process of clustering processing based on the density clustering method, the minimum number of points to form a dense area is set to 10, and the clustering metric index under different neighborhood radii is calculated respectively, and the neighborhood radius with the largest clustering metric index is selected as the final neighborhood radius, wherein the clustering metric index is:

[0022]

[0023] Where: O is the clustering metric index, d i,i is the average distance from the i-th data point to all other data points in the same first cluster, d i,e is the minimum average distance from the ith data point to any other data point in the first cluster, S bp is the set of data points with the navigation status of anchoring and mooring, x is the proportion of abnormal data points, B area is the parking area set, |B area | is the length of the parking area set.

[0024] The step S2 comprises:

[0025] Step S2-1: dividing all AIS data points into trajectories based on the ship code, and sorting the data points belonging to the same ship by time to obtain multiple first intermediate trajectories;

[0026] Step S2-2: dividing all first intermediate trajectories based on the navigation destination, and sorting the data points belonging to the same destination by time to obtain multiple second intermediate trajectories;

[0027] Step S2-3: for all the second intermediate trajectories, determine in turn whether the time interval between any two adjacent data points exceeds the first set threshold, and use all the two data points whose time interval exceeds the first set threshold as segmentation points to segment the second intermediate trajectories, thereby obtaining multiple potential valid trajectories;

[0028] Step S2-4: determine in turn whether the starting point and the end point of each potential valid trajectory are in the parking area. If the starting point and the end point of any potential valid trajectory are both in the parking area, the potential valid trajectory is regarded as a valid trajectory.

[0029] In step S3, for each valid trajectory, a data point with a navigation status of "navigating" is selected, and a process of selecting and obtaining a plurality of waypoints based on the ground speed, the heading difference and the speed difference includes:

[0030] Step S3-1: For each data point in the valid trajectory, select all data points whose navigation status is "navigating" as potential waypoints;

[0031] Step S3-2: Obtain the ground speed of each potential waypoint;

[0032] Step S3-3: selecting a potential waypoint whose ground speed is within a preconfigured ground speed threshold range as a first intermediate waypoint;

[0033] Step S3-4: Obtain the instantaneous speed of all first intermediate waypoints, the average speed and speed direction from the previous data point to the first intermediate waypoint, and the average speed and speed direction from the first intermediate waypoint to the next data point;

[0034] Step S3-5: taking the difference between the speed direction from the first intermediate waypoint to the next data point and the speed direction from the previous data point to the first intermediate waypoint as the heading difference;

[0035] Step S3-6: Obtaining a speed difference based on a difference between an average speed from the first intermediate waypoint to a subsequent data point and an average speed from a previous data point to the first intermediate waypoint;

[0036] Step S3-7: The first intermediate waypoint whose heading difference is greater than a preconfigured heading difference threshold and whose speed difference is within the preconfigured speed difference threshold range is taken as a waypoint.

[0037] The step S3-6 specifically includes:

[0038] Step S3-6-1: subtract the instantaneous speed of the first intermediate waypoint from the average speed from the first intermediate waypoint to the next data point to obtain an absolute value as the first speed value;

[0039] Step S3-6-1: subtract the instantaneous speed of the first intermediate waypoint from the average speed from the previous data point to the first intermediate waypoint to obtain an absolute value as the second speed value;

[0040] Step S3-6-2: The sum of the first speed value and the second speed value is taken as the speed difference.

[0041] In step S4, the DTW algorithm is used to perform trajectory fusion to obtain a unique arc between nodes, including:

[0042] Step S4-1: Select two navigation trajectories and use the DTW algorithm to fuse the trajectories to obtain a navigation trajectory;

[0043] Step S4-2: Determine whether there are two navigation tracks. If yes, return to step S4-1. Otherwise, take the only ship track as the arc between the two nodes.

[0044] A shipping traffic network extraction device based on ship movement data includes a memory, a processor, and a program stored in the memory, characterized in that the processor implements the above method when executing the program.

[0045] A storage medium stores a program, which implements the above method when executed.

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] 1. The coordinate data of the port area is used. The anchorage and berth range can more accurately reflect the actual activity area of ​​the ship. Secondly, the anchorage range can completely eliminate the influence of drifting AIS data.

[0048] 2. No complex algorithms are required. You only need to extract accurate ship berthing point data and determine whether the point is within the port range for accurate matching.

[0049] 3. Using the speed of the ship's AIS data, the time difference between adjacent data points, and the distance, the stop points can be extracted from the perspective of time and space. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 Schematic diagram of the shipping traffic network extraction framework based on ship movement data;

[0051] Figure 2 It is a schematic diagram of the main steps of the present invention;

[0052] Figure 3 Extracting concept maps for shipping networks based on ship movement data;

[0053] Figure 4 This is a schematic diagram of heading difference calculation;

[0054] Figure 5 Schematic diagram of trajectory DTW fusion;

[0055] Figure 6 Schematic diagram of the construction and model expression of the shipping transportation network. DETAILED DESCRIPTION

[0056] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0057] The present invention is a shipping network extraction framework and method based on ship movement data. Figure 3 As shown, characteristic data points such as anchorage points and turning points are identified from dynamic ship movement data, and ship data of different status categories are clustered to obtain the corresponding ship activity areas. Each area is regarded as a node in the network, and arcs between nodes are generated based on the ship trajectory to obtain the shipping traffic network.

[0058] According to the attached Figure 2 The process shown, the shipping traffic network extraction framework and process based on ship movement data includes 1 preprocessing module and 5 core modules, namely, berthing area recognition module, effective trajectory extraction module, waypoint area recognition module, network arc generation module and shipping network construction module.

[0059] A shipping traffic network extraction method based on ship movement data, such as Figure 1 and Figure 2 As shown, including:

[0060] Step S1: extracting data points with anchored and moored navigation status in AIS data as basic ship berthing point data, and clustering them using density-based clustering to obtain multiple berthing areas, including:

[0061] Step S1-1: Acquire AIS data and select data points with navigation status of anchoring and mooring as basic ship berthing point data;

[0062] After obtaining the original AIS data, unnecessary data columns are first filtered out and the data column contents used to construct the shipping traffic network are retained. The retained data column contents include ship MMSI, data timestamp, destination, positioning longitude and latitude, ship heading to the ground, ship speed to the ground, and navigation status.

[0063] Secondly, the longitude and latitude coordinates of the ship need to be converted. In order to facilitate the calculation of the spatial distance between the subsequent ship points, the longitude and latitude coordinates are converted to Mercator coordinates. Assuming that the longitude and latitude of the ship are lon and lat respectively, the Mercator abscissa x1 and ordinate y1 of the ship are:

[0064]

[0065] Where: R e is the radius of the earth, which can be 6381372.08 meters, and α represents the arc of latitude α=(lat*π) / 180

[0066] Step S1-2: performing clustering processing using a density-based clustering method to obtain a plurality of first clusters;

[0067] Specifically, after obtaining the ship berthing status data, the basic ship berthing point data is processed through the DBSCAN clustering algorithm to obtain different berthing areas. Since the number of ship berthing areas cannot be confirmed, the number of categories cannot be used as the input of the algorithm. At the same time, some ships choose to berth due to special circumstances, and the areas where they are located do not have the meaning of general berthing areas, that is, these berthing points are abnormal berthing points. Conventional clustering algorithms cannot obtain sufficiently accurate berthing areas. Therefore, in this application, the density-based clustering method is the DBSCAN algorithm. The DBSCAN algorithm does not require the input of the number of categories and can identify abnormal data. Therefore, the DBSCAN algorithm is selected for clustering to realize the identification of berthing areas.

[0068] The DBSCAN algorithm requires two parameters to be input, namely the neighborhood radius ε and the minimum number of points minPoints to form a dense area.

[0069] In the process of clustering based on density clustering, the minimum number of points to form a dense area is set to 10 to ensure that there are enough parking points in each parking area. Then, the clustering metric index under different neighborhood radii is calculated respectively. Specifically, the value range of ε is set to [1000m, 5000m], and the neighborhood radius with the largest clustering metric index is selected as the final neighborhood radius. The clustering metric index is:

[0070]

[0071] Where: O is the clustering metric index, d i,i is the average distance from the i-th data point to all other data points in the same first cluster, d i,e is the minimum average distance from the ith data point to any other data point in the first cluster, S bp is the set of data points with the navigation status of anchoring and mooring, x is the proportion of abnormal data points, B area is the parking area set, |B area | is the length of the parking area set.

[0072] Step S1-3: taking the center of each first cluster as the center of the first cluster, and taking the distance from the point with the largest distance from the center in the first cluster to the center as the radius of the first cluster, and constructing a parking area based on the center coordinates and radius of the first cluster;

[0073] Step S1-4: Group all parking areas into a parking area set

[0074] Step S2: Based on the AIS data points, the track is divided based on the ship code, the navigation destination and the time interval of adjacent data points to generate multiple valid tracks, including:

[0075] Step S2-1: Divide all AIS data points into trajectories based on the ship code, and sort the data points belonging to the same ship by time to obtain multiple first intermediate trajectories Where i represents different ships;

[0076] Step S2-2: For all first intermediate trajectories, divide the trajectories based on the navigation destination, and sort the data points belonging to the same destination by time to obtain multiple second intermediate trajectories

[0077] Step S2-3: for all the second intermediate trajectories, determine in turn whether the time interval between any two adjacent data points exceeds the first set threshold, and use all the two data points whose time interval exceeds the first set threshold as segmentation points to segment the second intermediate trajectories, thereby obtaining multiple potential valid trajectories;

[0078] Specifically, since there may still be temporary berthing points in the second intermediate trajectory or the data in the second intermediate trajectory is not continuous in time due to the long berthing time of the ship, in order to obtain the trajectory of the "berthing-sailing-berthing" mode, the time interval between every two adjacent points in the second intermediate trajectory is calculated, and 24 hours is taken as the time threshold. When the time interval exceeds 24 hours, the second intermediate trajectory is further segmented, and finally a potential valid trajectory is obtained.

[0079] Step S2-4: determine in turn whether the starting point and the end point of each potential valid trajectory are in the parking area. If the starting point and the end point of any potential valid trajectory are both in the parking area, specifically, they can be two different parking areas, then the potential valid trajectory is regarded as a valid trajectory.

[0080] Step S3: for each valid trajectory, select the data point whose navigation status is "navigating", select multiple waypoints based on ground speed, heading difference and speed difference, and cluster the waypoints obtained based on all valid trajectories using a density-based clustering method to obtain multiple stop areas, and merge the parking areas and stop areas to obtain a node set of the shipping traffic network;

[0081] In step S3, for each valid trajectory, a data point with a navigation status of "navigating" is selected, and a process of selecting and obtaining multiple waypoints based on the ground speed, the heading difference and the speed difference includes:

[0082] Step S3-1: For each data point in the valid trajectory, select all data points whose navigation status is "navigating" as potential waypoints;

[0083] Among them, the navigation specifically refers to the AIS data

[0084] Step S3-2: Obtain the ground speed of each potential waypoint;

[0085] Step S3-3: Select a potential waypoint whose ground speed is within a preconfigured ground speed threshold range as the first intermediate waypoint:

[0086]

[0087] Among them; SOG max and SOG min the upper and lower limits of the ground speed threshold range, respectively;

[0088] Step S3-4: Obtain the instantaneous speed of all first intermediate waypoints, the average speed and speed direction from the previous data point to the first intermediate waypoint, and the average speed and speed direction from the first intermediate waypoint to the next data point;

[0089] Step S3-5: The difference between the speed direction from the first intermediate waypoint to the next data point and the speed direction from the previous data point to the first intermediate waypoint is taken as the heading difference.

[0090] Step S3-6: Obtaining a speed difference based on the difference between the average speed from the first intermediate waypoint to the next data point and the average speed from the previous data point to the first intermediate waypoint, specifically comprising:

[0091] Step S3-6-1: average speed from the first intermediate waypoint to the next data point Subtract the instantaneous speed at the first intermediate waypoint The obtained absolute value is taken as the first speed value;

[0092] Step S3-6-1: Average speed from the previous data point to the first intermediate waypoint Subtract the instantaneous speed at the first intermediate waypoint The obtained absolute value is taken as the second speed value;

[0093] Step S3-6-2: Taking the sum of the first speed value and the second speed value as the speed difference

[0094]

[0095] Step S3-7: The heading difference is greater than the pre-configured heading difference threshold NCD. threshold , and the first intermediate waypoint whose speed difference is within the pre-configured speed difference threshold is used as the waypoint:

[0096]

[0097] Of which: SNSDmax and SNSD min The upper and lower limits of the speed difference threshold range are preconfigured respectively.

[0098] Specifically, the ship's track points of as well as Calculation

[0099] Ship Point Navigation status and ground speed Can be obtained from AIS data.

[0100] Consider the calculation of travel speed, from To the next data point Average speed Ground speed The difference is: ground speed Yes The instantaneous speed, From point To point The average driving speed can be calculated as follows.

[0101]

[0102] in, To calculate a point based on longitude and latitude and Point The sailing distance between

[0103] is the corresponding flight time difference. Therefore, we need to calculate according to the above formula and Then calculate separately With point The instantaneous speed difference and The difference in instantaneous speed is finally summed up to get

[0104] Secondly, point To point Direction It can also be calculated using longitude and latitude, as shown below.

[0105]

[0106] in, and Respectively indicate points and Point The longitude and latitude of , \ represents the remainder operation. and After that, different scenarios need to be considered for calculation In the attached Figure 4 (a) and attached Figure 4 In (b), All less than 180°, Figure 4 The value of CD in (a) is In the attached Figure 4 In (b) In the attached Figure 4 (c) and attached Figure 4 In (d), All are greater than 180°, Figure 4 The value of CD in (c) is In the attached Figure 4 In (d), In summary, The calculation of is shown below.

[0107]

[0108] After determining the calculation method of SNSD and NCD of the trajectory midpoint, the next step is to determine the threshold to determine the waypoint. threshold In order to clarify the correlation between speed change and heading change, and thus determine the corresponding threshold, SOG can be determined by calculating the speed change and heading change at each point based on the data collected from all navigation states, and using the data distribution density map. max , SOG min 、NCD threshold SNSD min and SNSD max After testing, the following parameter thresholds can be used: SOG max =15knots,SOG min =5knots,NCD threshold =10°,SNSD min =0m / s,andSNSD max =8m / s.

[0109] After determining the calculation method and threshold of the relevant variables, the data of all points in the navigation are screened to obtain the waypoints. After that, the DBSCAN algorithm and indicators are also used to construct the waypoint area, so as to obtain the set W of all the anchor point areas. area ,in Contains all waypoints in the area and corresponds to the center and radius of the area. area and B areaAfter fusion, we get the node set N of the shipping transportation network.

[0110] Step S4: Based on the valid trajectory, it is cyclically determined whether there is a navigation trajectory between the nodes. If there is a navigation trajectory, the ship trajectory is used as the arc between the two nodes. If there are multiple trajectories, the DTW algorithm is used to fuse the trajectories to obtain a unique arc between the nodes.

[0111] Among them, the DTW algorithm is used for trajectory fusion to obtain the unique arc between nodes, including:

[0112] Step S4-1: Select two navigation trajectories and use the DTW algorithm to fuse the trajectories to obtain a navigation trajectory;

[0113] Step S4-2: Determine whether there are two navigation tracks. If yes, return to step S4-1. Otherwise, take the only ship track as the arc between the two nodes.

[0114] Specifically, another construction level of the shipping transportation network is the arcs between nodes. When identifying waypoint areas and parking areas, the segmented trajectory used can be used to construct arcs between nodes. However, there may be multiple trajectories corresponding to a specific node. In order to obtain a unique line between node pairs to represent the arcs between nodes, the dynamic time warping DTW algorithm is used to fuse the trajectories. The DTW algorithm can find the optimal match between two time series. Given two time series Q = {q1, q2, …, q m} and C={c1,c2,…,c n}, we can construct an m×n matrix to correspond to the distance between every two corresponding elements of the matrix, thus obtaining a path W={w1,w2,…,w K}, kth th The elements of k =(a,b) k , where a and b represent the ordinal numbers of the matching points in the two trajectories.

[0115] Attached Figure 5 The DTW-based ship trajectory fusion process is shown. Through the DTW algorithm, the optimal point matching of the two is obtained, and then the average trajectory can be obtained from the arithmetic mean of the aligned positions. For each pair of nodes, multiple trajectories are fused into one trajectory by repeating this process to obtain the corresponding arcs between the nodes. Finally, the unique arc of each pair of connected nodes is obtained, that is, the set A of all arcs.

[0116] Step S5: construct a shipping transportation network based on the obtained node set and arcs between nodes.

[0117] Based on the obtained node set N and arc set A, a directed graph G = (N, A) of the shipping transportation network is constructed. In addition, in order to represent the connection between nodes, a node arc association matrix M is defined: the dimension is n*m, each row corresponds to a node and each column corresponds to an arc; the matrix (i, k) th The element a i,k With the i th nodes and the kth th Related, as shown below.

[0118]

[0119] Among them, speed represents the kth th Therefore, each column in M ​​contains two non-zero elements, 1*speed and -1*speed, and the sign represents the direction of the corresponding arc. Finally, the construction and model expression of the shipping transportation network are completed, as shown in the attached figure. Figure 6 shown.

[0120] If the above functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program code.

Claims

1. A shipping traffic network extraction method based on ship movement data, characterized in that: include: Step S1: extracting data points with navigation status of anchoring and mooring in AIS data as basic ship berthing point data, and performing clustering processing using density-based clustering to obtain multiple berthing areas; Step S2: According to the AIS data points, the track is divided based on the ship code, the navigation destination and the time interval of adjacent data points to generate multiple valid tracks; Step S3: for each valid trajectory, select the data point whose navigation status is "navigating", select multiple waypoints based on ground speed, heading difference and speed difference, and cluster the waypoints obtained based on all valid trajectories using a density-based clustering method to obtain multiple stop areas, and merge the parking areas and stop areas to obtain a node set of the shipping traffic network; Step S4: Based on the valid trajectory, it is cyclically determined whether there is a navigation trajectory between the nodes. If there is a navigation trajectory, the ship trajectory is used as the arc between the two nodes. If there are multiple trajectories, the DTW algorithm is used to fuse the trajectories to obtain a unique arc between the nodes. Step S5: construct a shipping transportation network based on the obtained node set and arcs between nodes.

2. A shipping traffic network extraction method based on ship movement data according to claim 1, characterized in that: The step S1 comprises: Step S1-1: Acquire AIS data and select data points with navigation status of anchoring and mooring as basic ship berthing point data; Step S1-2: performing clustering processing using a density-based clustering method to obtain a plurality of first clusters; Step S1-3: taking the center of each first cluster as the center of the first cluster, and taking the distance from the point with the largest distance from the center in the first cluster to the center as the radius of the first cluster, and constructing a parking area based on the center coordinates and radius of the first cluster; Step S1-4: All parking areas are grouped into a parking area set.

3. A shipping traffic network extraction method based on ship movement data according to claim 1 or 2, characterized in that: The density-based clustering method is the DBSCAN algorithm.

4. A shipping traffic network extraction method based on ship movement data according to claim 3, characterized in that: In the process of clustering processing based on the density clustering method, the minimum number of points to form a dense area is set to 10, and the clustering metric index under different neighborhood radii is calculated respectively, and the neighborhood radius with the largest clustering metric index is selected as the final neighborhood radius, wherein the clustering metric index is: Where: o is the clustering metric index, d i,i is the average distance from the i-th data point to all other data points in the same first cluster, d i,e is the minimum average distance from the ith data point to any other data point in the first cluster, S bp is the set of data points with the navigation status of anchoring and mooring, x is the proportion of abnormal data points, B area is the parking area set, |B area | is the length of the parking area set.

5. The method for extracting a shipping traffic network based on ship movement data according to claim 1 is characterized in that: The step S2 comprises: Step S2-1: dividing all AIS data points into trajectories based on the ship code, and sorting the data points belonging to the same ship by time to obtain multiple first intermediate trajectories; Step S2-2: dividing all first intermediate trajectories based on the navigation destination, and sorting the data points belonging to the same destination by time to obtain multiple second intermediate trajectories; Step S2-3: for all the second intermediate trajectories, determine in turn whether the time interval between any two adjacent data points exceeds the first set threshold, and use all the two data points whose time interval exceeds the first set threshold as segmentation points to segment the second intermediate trajectories, thereby obtaining multiple potential valid trajectories; Step S2-4: determine in turn whether the starting point and the end point of each potential valid trajectory are in the parking area. If the starting point and the end point of any potential valid trajectory are both in the parking area, the potential valid trajectory is regarded as a valid trajectory.

6. A shipping traffic network extraction method based on ship movement data according to claim 5, characterized in that: In step S3, for each valid trajectory, a data point with a navigation status of "navigating" is selected, and a process of selecting and obtaining a plurality of waypoints based on the ground speed, the heading difference and the speed difference includes: Step S3-1: For each data point in the valid trajectory, select all data points whose navigation status is "navigating" as potential waypoints; Step S3-2: Obtain the ground speed of each potential waypoint; Step S3-3: selecting a potential waypoint whose ground speed is within a preconfigured ground speed threshold range as a first intermediate waypoint; Step S3-4: Obtain the instantaneous speed of all first intermediate waypoints, the average speed and speed direction from the previous data point to the first intermediate waypoint, and the average speed and speed direction from the first intermediate waypoint to the next data point; Step S3-5: taking the difference between the speed direction from the first intermediate waypoint to the next data point and the speed direction from the previous data point to the first intermediate waypoint as the heading difference; Step S3-6: Obtaining a speed difference based on a difference between an average speed from the first intermediate waypoint to a subsequent data point and an average speed from a previous data point to the first intermediate waypoint; Step S3-7: The first intermediate waypoint whose heading difference is greater than a preconfigured heading difference threshold and whose speed difference is within the preconfigured speed difference threshold range is taken as a waypoint.

7. A method for extracting a shipping traffic network based on ship movement data according to claim 6, characterized in that: The step S3-6 specifically includes: Step S3-6-1: subtract the instantaneous speed of the first intermediate waypoint from the average speed from the first intermediate waypoint to the next data point to obtain an absolute value as the first speed value; Step S3-6-1: subtract the instantaneous speed of the first intermediate waypoint from the average speed from the previous data point to the first intermediate waypoint to obtain an absolute value as the second speed value; Step S3-6-2: The sum of the first speed value and the second speed value is taken as the speed difference.

8. The method for extracting a shipping traffic network based on ship movement data according to claim 5 is characterized in that: In step S4, the DTW algorithm is used to perform trajectory fusion to obtain a unique arc between nodes, including: Step S4-1: Select two navigation trajectories and use the DTW algorithm to fuse the trajectories to obtain a navigation trajectory; Step S4-2: Determine whether there are two navigation tracks. If yes, return to step S4-1. Otherwise, take the only ship track as the arc between the two nodes.

9. A shipping traffic network extraction device based on ship movement data, comprising a memory, a processor, and a program stored in the memory, characterized in that: When the processor executes the program, the method according to any one of claims 1 to 8 is implemented.

10. A storage medium having a program stored thereon, characterized in that: When the program is executed, the method according to any one of claims 1 to 8 is implemented.

Images