Processing method, system and device for extracting ship track line features and storage medium

Abstract

The application discloses a kind of processing methods, systems and devices for extracting ship track line features.The method obtains AIS dynamic data of the ship, and pre-processes the latitude and longitude data in the AIS dynamic data, thereby improving the efficiency of processing data in subsequent steps.The latitude and longitude data after pre-processing is converted into ship track image data, and the spatial features and auxiliary features of the track line are obtained according to the ship track image data.The model is trained by obtaining the spatial features and auxiliary features, which can make the simulation results of the model more consistent with the actual situation.According to the spatial features and auxiliary features, the classification results of the track line are generated by the ship track line feature extraction model, which can extract the features of the ship track line with higher accuracy.The application can be widely applied in the field of ship behavior pattern recognition technology.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of ship behavior pattern recognition, and in particular to a processing method, system and device for extracting ship track line features and a storage medium. BACKGROUND

[0002] A ship track line is a sequence of position and time records during ship navigation, and ship track line feature extraction as a specific study of trajectory data analysis can help maritime supervisors determine the channel to which the ship track line belongs and mine the navigation mode and behavior mode thereof.

[0003] In related technologies, a large number of machine learning algorithms have been proposed and applied to the study of ship track line feature extraction, such as support vector machine model algorithms, random forest algorithms, sparse representation algorithms, etc. These processing methods for extracting ship track line features have less use of spatial feature information of ship track lines, and the process of extracting features such as ship track line heading and speed is too cumbersome. SUMMARY

[0004] The present application aims to at least partially solve one of the technical problems existing in the prior art.

[0005] To this end, one object of the present application is to provide a processing method, system, device and medium for extracting ship track line features, which can extract the features of ship track lines with higher accuracy.

[0006] In order to achieve the above technical purpose, the technical solution adopted by the embodiments of the present application comprises:

[0007] On the one hand, the present application provides a processing method for extracting ship track line features, comprising the following steps:

[0008] Obtaining AIS dynamic data of a ship, the AIS dynamic data comprising latitude and longitude data;

[0009] Preprocessing the latitude and longitude data;

[0010] Converting the preprocessed latitude and longitude data into ship track image data;

[0011] Obtaining spatial features and auxiliary features of a track line according to the ship track image data;

[0012] Generating a classification result of the track line by a ship track line feature extraction model according to the spatial features and the auxiliary features.

[0013] Further, the step of preprocessing the latitude and longitude data specifically comprises:

[0014] unifying data formats of the ship trajectory image data;

[0015] eliminating abnormal data in the ship trajectory image data;

[0016] standardizing the ship trajectory image data.

[0017] Further, the step of converting the pre-processed latitude and longitude data into ship trajectory image data specifically includes:

[0018] determining a target rectangular region according to the latitude and longitude data;

[0019] determining a resolution of the target rectangular region through an aspect ratio of the target rectangular region;

[0020] determining a distance of latitude and longitude coordinate points according to the latitude and longitude data;

[0021] extracting the resolution and the distance of the latitude and longitude coordinate points through a residual network to obtain the ship trajectory image data.

[0022] Further, the step of obtaining spatial features and auxiliary features of the ship trajectory image data specifically includes:

[0023] obtaining the resolution and the distance of the latitude and longitude coordinate points in the ship trajectory image data as spatial features of the ship trajectory image data;

[0024] generating auxiliary features of the ship trajectory image data through a variational autoencoder.

[0025] Further, the training step of the ship trajectory feature extraction model includes:

[0026] setting a loss function;

[0027] obtaining trajectory training data;

[0028] inputting the trajectory training data into the ship trajectory feature extraction model for training.

[0029] Further, the training step further includes:

[0030] analyzing the training effect of the model through a data analysis visualization tool.

[0031] On the other hand, an embodiment of the present application proposes a processing system for extracting ship trajectory features, which includes:

[0032] a first module for obtaining AIS dynamic data of a target ship, the AIS dynamic data including latitude and longitude data;

[0033] The second module is configured to establish a ship track line feature extraction model.

[0034] The third module is configured to input the AIS dynamic data of the target ship into the ship track line feature extraction model to obtain a track line prediction result and a classification result of the target ship.

[0035] Further, the training step of establishing the ship track line feature extraction model comprises:

[0036] The latitude and longitude data is preprocessed.

[0037] The preprocessed latitude and longitude data is converted into ship track image data.

[0038] The ship track image data is input into the ship track line feature extraction model.

[0039] A loss function is set to train the ship track line feature extraction model.

[0040] In another aspect, an embodiment of the present application provides a processing device for extracting ship track line features, comprising:

[0041] At least one processor;

[0042] At least one memory for storing at least one program;

[0043] When the at least one program is executed by the at least one processor, the at least one processor implements the processing method for extracting ship track line features.

[0044] In another aspect, an embodiment of the present application provides a storage medium having processor-executable instructions stored therein, wherein the processor-executable instructions, when executed by a processor, are used to implement the processing method for extracting ship track line features.

[0045] The present application discloses a processing method for extracting ship track line features, which has the following advantages:

[0046] The present embodiment improves the efficiency of processing data in subsequent steps by obtaining AIS dynamic data of a ship and preprocessing latitude and longitude data in the AIS dynamic data. The preprocessed latitude and longitude data is converted into ship track image data, the spatial features and auxiliary features of the track line are obtained according to the ship track image data, the model is trained by obtaining the spatial features and auxiliary features, and the simulation result of the model is more in line with the actual situation. According to the spatial features and auxiliary features, the classification result of the track line is generated by the ship track line feature extraction model, and the features of the ship track line can be extracted with higher accuracy. BRIEF DESCRIPTION OF DRAWINGS

[0047] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following introduces the drawings of the related technical solutions in the embodiments of the present application or the prior art. It should be understood that the drawings in the following introduction are only for the convenience of clearly describing part of the embodiments of the technical solutions of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the premise that there is no creative labor.

[0048] Figure 1 A flowchart of a processing method for extracting ship track line features provided by the embodiment of the present application;

[0049] Figure 2 A module schematic diagram of a ship track line analysis system provided by the embodiment of the present application;

[0050] Figure 3 A structure schematic diagram of a deep convolutional neural network provided by the embodiment of the present application;

[0051] Figure 4 A structure schematic diagram of a processing system for extracting ship track line features provided by the embodiment of the present application;

[0052] Figure 5 A structure schematic diagram of a processing device for extracting ship track line features provided by the embodiment of the present application. DETAILED DESCRIPTION

[0053] This part will describe the specific embodiments of the present application in detail, and the preferred embodiments of the present application are shown in the drawings. The drawings serve to supplement the description in the text part of the specification, so that people can intuitively and visually understand each technical feature and the overall technical solution of the present application. However, it cannot be understood as a limitation on the protection scope of the present application.

[0054] In the description of the embodiments of the present application, one or more is meant to be one or more, and more than two is meant to be two or more. Greater than, less than, more than, etc. are understood as not including the number. Above, below, within, etc. are understood as including the number. "At least one" means one or more. "At least one of the following" and the like means any combination of these items, including single or multiple items. If "first", "second", etc. are described, it is only for the purpose of distinguishing technical features, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features or the order of indicated technical features.

[0055] It should be noted that the terms such as setting, installing, connecting and the like in the embodiments of the present application should be understood in a broad sense, and the skilled in the art can determine the specific meaning of the above terms in the embodiments of the present application in combination with the specific content of the technical solutions. For example, the term "connection" can be mechanical connection, electrical connection or can communicate with each other; it can be directly connected, or indirectly connected through an intermediate medium.

[0056] In the description of the embodiments of the present application, the description of the terms "one embodiment", "another embodiment" or "some embodiments", "in the above embodiment" and the like means that the specific features, structures, materials or characteristics described in combination with the embodiments or examples are contained in at least two embodiments or embodiments of the present disclosure. In the present disclosure, the illustrative description of the above terms does not necessarily refer to the same example or embodiment. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or embodiments in a suitable manner.

[0057] It should be noted that the technical features involved in each of the embodiments of the present application described below can be combined with each other as long as there is no conflict.

[0058] Track is the trajectory of a ship or a water aircraft when sailing. The ship will be affected and disturbed by various factors during sailing, resulting in sailing errors, so the track is generally not a straight line. The line connecting adjacent estimated ship positions is called estimated track, and the line connecting adjacent observed ship positions is called actual track.

[0059] Ship track line is a sequence of position and time records during ship sailing. Ship track line feature extraction, as a specific study of trajectory data analysis, can help maritime supervisors determine the channel to which the ship track line belongs, and mine the sailing mode and behavior mode.

[0060] At present, a large number of machine learning algorithms are used in the research of ship track line feature extraction, such as support vector machine model algorithm, random forest algorithm, sparse representation algorithm, etc. However, these ship track line feature extraction methods have the problem of low utilization rate of spatial feature information of ship track line, and the extraction process of these methods is too cumbersome and inefficient.

[0061] To this end, the application provides a processing method, system and device for extracting ship track line features and a medium. The method comprises the following steps: acquiring AIS dynamic data of a target ship; establishing a ship track line feature extraction model; training the ship track line feature extraction model; inputting the AIS dynamic data of the target ship into the trained ship track line feature extraction model; and obtaining a track line prediction result and a classification result of the target ship. The method can extract the features of the ship track line with higher accuracy.

[0062] Referring to Figure 1 , the embodiment of the application provides a processing method for extracting ship track line features. The method can be applied to a ship track line analysis system as shown in the figure. Figure 2 The AIS data acquisition module 201 is used for AIS data, the memory 203 is used for storing ship track line images, latitude and longitude and heading angle data, and the processor 202 completes the processing method for extracting ship track line features by calling the data in the memory 203.

[0063] Referring to Figure 2 , Figure 2 is a module schematic diagram of a ship track line analysis system related to the embodiments of the application. In the embodiments of the application, the data aggregation system can include a communication bus, an AIS data acquisition module 201, a processor 202 (for example, a central processing unit, CPU) and a memory 203. The communication bus is used to realize the connection and communication between the components; the input port is used for data input; the output port is used for data output, and the memory 203 can be a high-speed RAM memory or a stable memory (for example, a disk memory), and the memory 203 can optionally be a storage device independent of the aforementioned processor 202. Those skilled in the art can understand that the hardware structure shown in the figure does not constitute a limitation on the application, and can include more or fewer components than the figure, or combine certain components, or different component arrangements. Figure 2

[0064] The memory 203 as a readable storage medium can include an operating system, an application program module and a control program of a distributed platform. In the embodiment of the application, the AIS data acquisition module 201 is mainly used for acquiring AIS data and transmitting the AIS data to the processor 202; and the processor 202 can call the control program of the ship track line analysis stored in the memory 203 and execute the processing method for extracting ship track line features provided by the embodiments of the application. Figure 2

[0065] ​​It should be noted that only a part of the modules of the ship track line analysis system is exemplarily given in the embodiments of the present application, and the ship track line analysis system can further include other components included in other existing ship track line analysis systems to realize corresponding functions, and the specific implementation is not limited.

[0066] Based on Figure 2 As shown in the ship track line analysis system, Figure 1 The embodiments provide a processing method for extracting ship track line features, including but not limited to steps S101, S102, S103, S104 and S105:

[0067] S101, acquiring AIS dynamic data of the ship, the AIS dynamic data including latitude and longitude data.

[0068] It should be noted that the AIS dynamic data refers to data obtained through AIS technology. Satellite AIS is a ship positioning technology. The low-orbit satellite receives AIS message information sent by the ship, and the satellite forwards the received and decoded AIS message information to the corresponding earth station, so that the land management agency can master the relevant dynamic information of the ship, and realize the monitoring of the ship sailing in the open sea. The AIS dynamic data in step S101 can be acquired through an AIS data acquisition module. The AIS data acquisition module 201 can be connected with the ship integrated navigation system through network communication, so as to read the AIS dynamic data therein. The AIS data acquisition module 201 also accesses the ship integrated navigation system through an API interface, so as to acquire the AIS dynamic data. The specific acquisition mode can be determined by those skilled in the art according to the actual situation. The AIS dynamic data includes latitude and longitude data of the track line.

[0069] S102, pre-processing the latitude and longitude data.

[0070] Optionally, step S102 can further include the following steps:

[0071] unifying the data format of the ship track image data;

[0072] eliminating abnormal data in the ship track image data;

[0073] standardizing the ship track image data.

[0074] In step S102, the acquired AIS dynamic data needs to be preprocessed, that is, the data format of the ship trajectory image data needs to be unified; the abnormal data in the ship trajectory image data needs to be eliminated; and the ship trajectory image data needs to be standardized. Because AIS dynamic data has various formats, processing various types of data separately will consume too much time and effort of users when processing a large amount of data, so unifying the data can effectively improve the efficiency of data processing. Then, the AIS dynamic data may contain abnormal data, and the AIS dynamic data needs to be cleaned to eliminate abnormal data in the ship trajectory image data. Data cleaning refers to the process of re-examining and checking data, aiming to delete repeated information, correct errors, and provide data consistency. This is a process of finding and correcting identifiable errors in data files, including checking data consistency, processing invalid values and missing values, etc. Then, the ship trajectory image data is standardized, that is, the maximum and minimum values of the detected water area longitude and latitude are taken as the reference, and the data is normalized by the maximum and minimum values, so that the value range of all data is between [0, 1].

[0075] S103, converting the preprocessed longitude and latitude data into ship trajectory image data.

[0076] Optionally, step S103 can further include the following steps:

[0077] determining a target rectangular region according to the longitude and latitude data;

[0078] determining the resolution of the target rectangular region by the aspect ratio of the target rectangular region;

[0079] determining the distance of the longitude and latitude coordinate points according to the longitude and latitude data;

[0080] extracting the resolution and the distance of the longitude and latitude coordinate points by a residual network to obtain the ship trajectory image data.

[0081] In step S103, the track latitude and longitude data needs to be converted into ship track image data, so as to extract the spatial features of the ship track line. First, the water area where the experimental data is located is regarded as a rectangular area, and the length and width of the rectangular area are calculated to obtain the aspect ratio to determine the image resolution. Considering that the higher the image resolution, the higher the required calculation cost, and the lower the calculation efficiency of the deep convolutional neural network, the resolution is set to 112*72, so that the latitude value of [49.06, 48.90] is mapped to the range of pixel points [0, 71], and the longitude value of [-123.42, -123.03] is mapped to the range of pixel points [0, 111]. Then, the distance between two latitude and longitude coordinate points is calculated using the Haversine formula. Then, the spatial features of the track line are extracted by ResNet50. The Haversine formula is also called the versine formula, and the distance between two points can be calculated according to the latitude and longitude through the versine formula.

[0082] S104, obtaining the spatial features and auxiliary features of the track line according to the ship track image data.

[0083] Optionally, step S104 can further include the following steps:

[0084] obtaining the resolution in the ship track image data and the distance of the latitude and longitude coordinate points as the spatial features of the track line;

[0085] generating the track line auxiliary features by a variational autoencoder according to the ship track image data.

[0086] The ship track image data containing the resolution and the distance of the latitude and longitude coordinate points can be obtained through the above S103 step, the resolution and the distance of the latitude and longitude coordinate points are taken as the spatial features of the track line, and the heading angle average change rate and the maximum and minimum curvatures are extracted from the track line image data as the auxiliary features of the track line by a variational autoencoder, the hyperparameters and the loss function of the model are set, and the model is unsupervised trained.

[0087] S105, generating the classification results of the track line according to the spatial features and the auxiliary features by a ship track line feature extraction model.

[0088] The training of the ship track line feature extraction model includes the following steps:

[0089] setting a loss function;

[0090] obtaining track training data;

[0091] inputting the track training data into the ship track line feature extraction model for training.

[0092] Specifically, the track line image data is input into the variational autoencoder, and the average change rate of the heading angle, the maximum and minimum curvatures can be obtained. In combination with the model task, the loss function of the model is set as follows (in the model training process, the smaller the value of the loss function is, the better, and the global minimum value of the loss function is taken as the training end condition):

[0093]

[0094] Where x is the AIS dynamic data that needs to be subjected to trajectory feature extraction; z is the hidden variable corresponding to x, in this example, the above certain water area of the Three Gorges of the Yangtze River is taken as an example, which is a vector with a dimension of 16; is the probability distribution possessed by the original AIS dynamic data, which exists objectively but cannot be directly obtained; q(x|z) is a normal distribution with a mean of G(x) and a constant variance, i.e. the output of the decoder; p(z|x) is a normal distribution with a mean of μ(x) and a variance of δ 2 (x) and δ 2 (x) are obtained through a neural network, i.e. the output of the encoder; q(z) is a standard normal distribution.

[0095] IE z～p(z|x) is the expectation of the expression under the condition that the distribution of z is p(z|x). The Adam optimization algorithm is used to solve the minimum value of the loss function.

[0096] IE z～p(z|x) [-logq(x|z)] indicates that the closer the x input into the encoder and the x' output by the decoder are, the better, which is equivalent to a regularization term, which makes all p(z|x) look like a standard normal distribution, and guarantees the generation ability of the model.

[0097] Then, the trajectory training data is obtained to train the model. The trajectory training data can be AIS data collected by the AIS data acquisition module 201 as training data. Specifically, the ship track line feature extraction model based on ResNet50 is as shown in Figure 3As shown, the structure is stacked by five convolutional blocks in turn, each of which contains a residual network substructure, and the residual network substructures in different convolutional blocks have different numbers of convolutional kernels. The input layer dimension parameter of the network model can be set to 112*72*3, the Mini Batch Size can be set to 64, and the output layer category can be set to 5. The iteration number of the model can be set to 20, and as the iteration number increases, the loss curve of the model shows a convergent trend, and there is a small amplitude fluctuation between the 4th and 8th iterations, but finally converges to about 0.04, obtaining a trained model. The training data for training the image feature extraction model is executed again in the trained model, and the data after all encoding and decoding of the variational autoencoder is obtained. As the iteration number increases, the loss curve of the ship track line image feature extraction model shows a convergent trend. The training result shows that the model has learned the data characteristics of the ship track line image, and the model training achieves the expected effect.

[0098] Optionally, the processing method for extracting ship track line features of the embodiment of the application can further include the following steps:

[0099] The training effect of the model is analyzed by a data analysis visualization tool.

[0100] In this embodiment, the Tensorboard data analysis visualization tool in TensorFlow is used to analyze the model training effect. TensorFlow is a symbolic mathematical system based on data flow programming, which is widely used in the programming implementation of various machine learning algorithms. TensorFlow has a multi-level structure and can be deployed on various servers, PC terminals and web pages and supports GPU and TPU high-performance numerical calculation.

[0101] Based on the foregoing description, taking the ship track line analysis system as an example, as shown in Figure 2 The processing method for extracting ship track line features of the embodiment of the application specifically includes the following steps:

[0102] S101, acquiring AIS dynamic data of a ship, the AIS dynamic data including latitude and longitude data.

[0103] S102, preprocessing the latitude and longitude data.

[0104] S103, converting the preprocessed latitude and longitude data into ship track image data.

[0105] S104, acquiring spatial features and auxiliary features of a track line according to the ship track image data.

[0106] S105, generating a classification result of the track line according to the spatial features and the auxiliary features through the ship track line feature extraction model.

[0107] Exemplarily, first, the AIS dynamic data of the ship needs to be acquired, the AIS dynamic data including latitude and longitude data, the AIS dynamic data being acquired through an AIS data acquisition module, the AIS data acquisition module 201 being connected with a ship integrated navigation system through network communication to read the AIS dynamic data therein, and the AIS data acquisition module 201 also accessing the ship integrated navigation system through an API interface to acquire the AIS dynamic data. Next, because the AIS dynamic data is not unified in data format and contains abnormal data, the latitude and longitude data needs to be preprocessed, that is, the data format of the ship track image data needs to be unified, the abnormal data in the ship track image data needs to be eliminated, and the ship track image data needs to be standardized. Then, the preprocessed latitude and longitude data is converted into ship track image data, first, the water area where the experimental data is located is regarded as a rectangular region, the length and width of the rectangular region are calculated to obtain an aspect ratio to determine the image resolution. Next, the distance between two latitude and longitude coordinate points is calculated using the Haversine formula. Then, the spatial features of the track line are extracted through ResNet50. After the ship track image data is acquired, the auxiliary features of the track line are generated through a variational autoencoder according to the ship track image data, the heading angle average change rate, the maximum and minimum curvatures are obtained by inputting the track line image data into the variational autoencoder. After the spatial features and the auxiliary features are acquired, the classification result of the track line is generated through the ship track line feature extraction model.

[0108] From the above, by acquiring the AIS dynamic data of the target ship, then establishing the ship track line feature extraction model, training the ship track line feature extraction model, and inputting the AIS dynamic data of the target ship into the trained ship track line feature extraction model, the track line prediction result and the classification result of the target ship are obtained. Through this method, the features of the ship track line can be extracted with higher accuracy. In addition, the model proposed by the present application achieves an accuracy of 98.72%, which is higher than the accuracy of 93.53% of the fully connected neural network model and the accuracy of 91.73% of the model support vector machine.

[0109] With reference to Figure 4 The processing system for extracting ship track line features proposed in the embodiment of the present application comprises:

[0110] The first module 401 is configured to acquire the AIS dynamic data of the ship, the AIS dynamic data including latitude and longitude data.

[0111] The second module 402 is configured to preprocess the latitude and longitude data.

[0112] The third module 403 is configured to convert the preprocessed latitude and longitude data into ship trajectory image data.

[0113] The fourth module 404 is configured to acquire spatial features and auxiliary features of the track line according to the ship trajectory image data.

[0114] The fifth module 405 is configured to generate a classification result of the track line by a ship track line feature extraction model according to the spatial features and the auxiliary features.

[0115] The above method embodiments are applicable to the system embodiments, the system embodiments achieve the same functions as the above method embodiments, and achieve the same beneficial effects as the above method embodiments.

[0116] With reference to Figure 5 The embodiment of the present application provides a processing device for extracting ship track line features, which comprises:

[0117] at least one processor 501;

[0118] at least one memory 502 configured to store at least one program;

[0119] When the at least one program is executed by the at least one processor 501, the at least one processor 501 is caused to implement Figure 1 the processing method for extracting ship track line features shown in the figure.

[0120] The above method embodiments are applicable to the device embodiments, the device embodiments achieve the same functions as the above method embodiments, and achieve the same beneficial effects as the above method embodiments.

[0121] The embodiment of the present application further provides a storage medium, wherein the storage medium stores processor-executable instructions, and the processor-executable instructions are used for implementing Figure 1 the processing method for extracting ship track line features shown in the figure when executed by a processor.

[0122] The above is a specific description of the preferred embodiment of the present application, but the present application is not limited to the above-mentioned embodiments, and those skilled in the art can make various equivalent modifications or replacements without departing from the spirit of the present application, and these equivalent modifications or replacements are all included in the scope defined by the claims of the present application.

Claims

1. A processing method for extracting features of a ship's track line, characterized in that, Includes the following steps: Acquire the ship's AIS dynamic data, which includes latitude and longitude data; The latitude and longitude data are preprocessed; The preprocessed latitude and longitude data is converted into ship trajectory image data, including: The target rectangular area is determined based on the latitude and longitude data; The resolution of the target rectangular region is determined by the aspect ratio of the target rectangular region. The distance between the latitude and longitude coordinates is determined based on the latitude and longitude data; The resolution and the distance between the latitude and longitude coordinates are extracted using a residual network to obtain the ship trajectory image data; The distance between two points can be calculated using the semi-versus formula based on latitude and longitude. Based on the ship trajectory image data, spatial features and auxiliary features of the track line are obtained, including: The resolution and the distance between the latitude and longitude coordinates in the ship trajectory image data are used as the spatial features of the track line. Based on the ship trajectory image data, a variational autoencoder is used to generate trackline auxiliary features, including: The variational autoencoder is used to extract the average rate of change of heading angle, maximum and minimum curvature as auxiliary features of the trajectory from the ship trajectory image data. Based on the spatial features and the auxiliary features, a classification result for the ship track line is generated using a ship track line feature extraction model, including: Set the loss function; The loss function: ； Where x is the AIS dynamic data for which trajectory feature extraction is required; z is the latent variable corresponding to x; This is the output of the decoder; This is the output of the encoder; It follows a standard normal distribution; For a z-distribution of Find the expectation of the expression under the given conditions.

2. The processing method for extracting ship track features according to claim 1, characterized in that, The step of preprocessing the latitude and longitude data specifically includes: Unify the data format of the ship trajectory image data; Remove abnormal data from the ship trajectory image data; The ship trajectory image data is standardized.

3. The processing method for extracting ship track features according to claim 1, characterized in that, The training steps of the ship track feature extraction model include: Obtain trajectory training data; The trajectory training data is input into the ship track feature extraction model for training.

4. The processing method for extracting ship track features according to claim 3, characterized in that, The training steps also include: The training performance of the model was analyzed using data analysis and visualization tools.

5. A processing system for extracting ship track features, used to implement the method as described in any one of claims 1 to 4, characterized in that, include: The first module is used to acquire the ship's AIS dynamic data, which includes latitude and longitude data; The second module is used for preprocessing the latitude and longitude data; The third module is used to convert the preprocessed latitude and longitude data into ship trajectory image data, including: The target rectangular area is determined based on the latitude and longitude data; The resolution of the target rectangular region is determined by the aspect ratio of the target rectangular region. The distance between the latitude and longitude coordinates is determined based on the latitude and longitude data; The resolution and the distance between the latitude and longitude coordinates are extracted using a residual network to obtain the ship trajectory image data; The distance between two points can be calculated using the semi-versus formula based on latitude and longitude. The fourth module is used to obtain the spatial features and auxiliary features of the track line based on the ship trajectory image data, including: The resolution and the distance between the latitude and longitude coordinates in the ship trajectory image data are used as the spatial features of the track line. Based on the ship trajectory image data, a variational autoencoder is used to generate trackline auxiliary features, including: The variational autoencoder is used to extract the average rate of change of heading angle, maximum and minimum curvature as auxiliary features of the trajectory from the ship trajectory image data. The fifth module is used to generate a classification result of the ship track line based on the spatial features and the auxiliary features using a ship track line feature extraction model, including: Set the loss function; The loss function: ； Where x is the AIS dynamic data for which trajectory feature extraction is required; z is the latent variable corresponding to x; This is the output of the decoder; This is the output of the encoder; It follows a standard normal distribution; For a z-distribution of Find the expectation of the expression under the given conditions.

6. The processing system for extracting ship track features according to claim 5, characterized in that, The training steps for establishing a ship track feature extraction model include: Obtain trajectory training data; The trajectory training data is input into the ship track feature extraction model for training.

7. A processing device for extracting features of ship tracks, characterized in that, include: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the processing method for extracting ship track features as described in any one of claims 1-4.

8. A computer-readable storage medium storing processor-executable instructions, characterized in that, The processor-executable instructions, when executed by the processor, are used to implement the processing method for extracting ship track features as described in any one of claims 1-4.

Images