Navigation bridge area ship trajectory prediction method based on graph neural network

Abstract

The invention relates to a navigation bridge area ship trajectory prediction method based on a graph neural network, and the method comprises the following steps: S1, setting a navigation bridge area ship monitoring field, and obtaining ship information and scene information; the monitoring field comprises an image information acquisition module, a ship basic information acquisition module and an environment information acquisition module; s2, establishing a navigation bridge area ship prediction model based on a graph neural network; the historical feature data is divided into two parts: a single object feature and a traffic flow feature; constructing a trajectory prediction model of a single ship through a graph neural network based on the features of the single object; generating a random ship flow simulation result, merging the random ship flow simulation result into a graph neural network as a new feature, and combining the new feature with a single ship trajectory prediction model to form a ship cluster trajectory prediction model; and S22, model prediction and training. Compared with the prior art, the method has the advantages that multi-source data fusion is realized, and the prediction accuracy and reliability are improved; the model design is targeted; the adaptability is high, and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of ship trajectory prediction, and particularly to a ship trajectory prediction method for a navigable bridge area based on a graph neural network. Background Art

[0002] In recent years, the global shipping industry has developed rapidly. The increase in ship-bridge collision accidents has highlighted the urgent need for a powerful ship trajectory prediction method in accident prevention. However, there are few ship trajectory prediction methods for navigable bridge areas at present. Traditional ship trajectory prediction methods only targeting ships do not fully consider the environmental characteristics and bridge structure characteristics of the bridge area. Summary of the Invention

[0003] The purpose of the present invention is to overcome the above-mentioned defects existing in the prior art and provide a ship trajectory prediction method for a navigable bridge area based on a graph neural network, which realizes the ship trajectory prediction considering the global information of the navigable bridge area and has strong portability and important reference significance for the active ship collision prevention warning of bridges.

[0004] The present invention provides a ship trajectory prediction method for a navigable bridge area based on a graph neural network, including the following steps:

[0005] S1: Set up a ship monitoring field in the navigable bridge area to obtain ship information and scene information; the monitoring field includes: an image information acquisition module, a ship basic information acquisition module, and an environmental information acquisition module;

[0006] The image information acquisition module includes a main bridge camera and a shoreline auxiliary camera. The image information acquisition module obtains information by using the method of target instance segmentation plus target tracking; the main bridge camera is the main way to obtain the appearance and trajectory characteristics of the ship, and has high requirements for the field of view, and a camera with better performance can be considered; considering the cost, the shoreline auxiliary camera is an auxiliary camera, mainly used to supplement the depth information of the image; the entire image information acquisition module is required to achieve full coverage of the selected navigable bridge area.

[0007] The ship basic information acquisition module includes an AIS data receiver and a radar; the static information, dynamic information, voyage information, and safety information of the ship are obtained through the AIS; the radar is used to make up for the defect of low recognition accuracy of the image information acquisition module under low visibility conditions;

[0008] The environmental information acquisition module includes a water level gauge and a wind speed sensor. The real-time water level and flow rate are obtained through the water level gauge, and the real-time wind speed is obtained through the wind speed sensor.

[0009] S2: Establish a ship prediction model for the navigable bridge area based on the graph neural network:

[0010] S21: Data integration and model establishment: Using the entire monitoring system, various types of characteristic information of the navigable bridge area can be obtained, including environmental characteristics, bridge site characteristics, and ship characteristics. Integrate these multi-source data to form historical characteristic data. The historical characteristic data is divided into two parts:

[0011] Single object characteristics: Cover the static and dynamic characteristics of a single ship, as well as characteristics such as the span of the navigable bridge that do not consider the coupling relationship between similar objects. Based on these single object characteristics, a trajectory prediction model for a single ship is constructed through a graph neural network.

[0012] Traffic flow characteristics: Used to consider the coupling relationship between ships. Given the complexity of ship traffic flow characteristics, a multi-level Monte Carlo sampling method is used to generate random ship flow simulation results, and these results are incorporated as new features into the graph neural network and combined with the single ship trajectory prediction model to form a ship cluster trajectory prediction model.

[0013] S22: Model prediction and training: The data collected by the monitoring field at time t is used as features and input into the model. Through continuous cyclic operations, the prediction of the subsequent trajectory of the ship can be achieved. At the same time, the predicted data can be used as new input for model training to achieve a data-driven effect.

[0014] Furthermore, in S1, as the internationally unified unique ship coding, MMSI can be defined as the label of the subsequent target detection model. This module identifies the basic information of the ship and serves as the label in the process of target instance segmentation plus target tracking to facilitate subsequent data integration.

[0015] Furthermore, in S1, the information collected by the environmental information collection module also includes bridge and bridge site information, including bridge navigation hole information and shoreline engineering.

[0016] Furthermore, in S21, when integrating multi-source data, data fusion technologies, including Kalman filtering and Bayesian fusion, are used to fuse data from different sources to improve the accuracy and reliability of the data.

[0017] Furthermore, in S21, feature selection and dimensionality reduction processing are performed on the historical characteristic data to remove redundant features and retain features that have an important impact on trajectory prediction. Principal component analysis (PCA) and linear discriminant analysis (LDA) methods are used for feature selection and dimensionality reduction.

[0018] Furthermore, in S21, an attention mechanism is introduced into the graph neural network to enable the model to pay more attention to features and nodes that are important for trajectory prediction. The attention mechanism can adaptively allocate weights to different features and nodes, improving the prediction accuracy of the model.

[0019] Further, in S21, a reinforcement learning algorithm is used to optimize the model. Reinforcement learning can continuously adjust the model's parameters through interaction with the environment to achieve optimal trajectory prediction. The ship's navigation trajectory is regarded as the state, and the prediction accuracy is regarded as the reward, and the model is continuously optimized through the reinforcement learning algorithm.

[0020] Further, in S22, in addition to using the prediction data for model training, the prediction results can also be compared and analyzed with the actual navigation trajectory to evaluate the prediction accuracy of the model. The mean square error (MSE) and mean absolute error (MAE) metrics are used to evaluate the performance of the model.

[0021] Further, in S21, the data integration of the collected information is achieved through the following methods:

[0022] S211: Identify the data sources and types;

[0023] S212: Associate with the ship as the core. Using the ship's MMSI code as the key identifier, associate the basic ship information obtained by the AIS data receiver with the ship's appearance and trajectory information obtained through object instance segmentation and object tracking in the image information acquisition module, so that different types of data of the same ship can be corresponding; for radar data, it can also be matched and associated with the ships in the AIS and image information according to information such as the ship's position. Especially when the visibility is low, use radar data to supplement and correct the deficiencies in the ship position information in other data.

[0024] S213: Integrate the environmental information and ship information. Integrate the water level, flow rate, wind speed, and bridge and bridge site information obtained by the environmental information acquisition module with the relevant information of the ship; combine the real-time water level and flow rate information with the ship's navigation dynamic information to analyze the impact of water flow on the ship's navigation trajectory; associate the wind speed information with the characteristics such as the ship's stability to consider the effect of wind on the ship's navigation.

[0025] S214: Construct historical feature data;

[0026] S215: Data preprocessing and normalization.

[0027] Further, in S22, the specific process of constructing the model is as follows:

[0028] S221: Determine the nodes and edges of the graph;

[0029] S222: Select the graph neural network architecture: Select the GCN graph convolutional neural network or GAT graph attention network or GraphSAGE;

[0030] S223: Design the layer structure of the model:

[0031] Input layer: The initial feature vectors of the nodes are input into the model, including various attribute features of the ship, environmental parameters, bridge and bridge site information. These features undergo preliminary linear transformation or embedding mapping through the input layer and are converted into low-dimensional vector representations suitable for graph neural network processing.

[0032] Hidden layer: Multiple graph neural network layers are stacked, such as GCN layer, GAT layer or GraphSAGE layer, as hidden layers. In each layer, by aggregating the features of neighboring nodes and updating the features of the current node, high-order relationships and complex patterns in the graph data are gradually extracted. The number of hidden layers and neurons needs to be adjusted according to specific problems and data characteristics, and generally the optimal structure is determined through experiments.

[0033] Output layer: The output layer is designed according to the goal of the prediction task. If predicting the trajectory of the ship, the output layer can be a fully connected layer, outputting the position coordinates, speed, and heading information of the ship at a future time step. If predicting events such as whether a ship will collide, the output layer can be a binary classification layer, outputting a probability value indicating the possibility of a collision.

[0034] S224: Define the loss function and optimizer.

[0035] S225: Model training and adjustment.

[0036] Compared with the prior art, the present invention has the following advantages:

[0037] (1) Multi-source data fusion, improving prediction accuracy and reliability: This method comprehensively utilizes multi-source data such as image information, basic ship information, and environmental information, and can comprehensively describe the ship navigation situation in the navigable bridge area from multiple perspectives, providing rich information for trajectory prediction and improving the accuracy and reliability of prediction.

[0038] (2) Targeted model design, strong adaptability: Different trajectory prediction models are designed for single ships and ship clusters respectively, considering the coupling relationship between ships and the complexity of traffic flow characteristics, enabling the model to better adapt to different scenarios and requirements.

[0039] (3) Data-driven learning mechanism: The predicted data is used as new input for model training, forming a data-driven learning mechanism, which can continuously optimize the performance of the model and improve the prediction accuracy. Description of the Drawings

[0040] Figure 1 It is a flowchart of the ship trajectory prediction method in the navigable bridge area based on graph neural network.

[0041] Figure 2 It is a layout diagram of the monitoring field.

[0042] Figure 3 It is the architecture diagram of the graph neural network at time t. Specific implementation manners

[0043] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. In the technical solution, features such as component models, material names, connection structures, control methods, algorithms, etc. that are not clearly described shall be regarded as common technical features disclosed in the prior art.

[0044] Embodiment 1

[0045] The present invention provides a method for predicting the trajectory of ships in a navigable bridge area based on a graph neural network. As Figure 1 shown, it includes the following steps:

[0046] S1: As Figure 2 shown, set up a ship monitoring field in the navigable bridge area to obtain ship information and scene information; the monitoring field includes: an image information collection module, a ship basic information collection module, and an environmental information collection module;

[0047] The image information collection module includes a main bridge camera and a shoreline secondary camera. The image information collection module obtains information by using object instance segmentation plus object tracking; the main bridge camera is the main way to obtain the appearance and trajectory features of ships and has high requirements for the field of view. A camera with good performance can be considered; considering the cost, the shoreline secondary camera is an auxiliary camera, mainly used to supplement the depth information of the image; the entire image information collection module is required to achieve full coverage of the selected navigable bridge area.

[0048] The ship basic information collection module includes an AIS data receiver and a radar; static information, dynamic information, voyage information, and safety information of ships are obtained through AIS; the radar is used to make up for the defect of low recognition accuracy of the image information collection module under low visibility conditions;

[0049] The environmental information collection module includes a water level gauge and a wind speed sensor. The real-time water level and flow rate are obtained through the water level gauge, and the real-time wind speed is obtained through the wind speed sensor.

[0050] S2: Establish a ship prediction model for the navigable bridge area based on the graph neural network:

[0051] S21: Data integration and model establishment: Using the entire monitoring system, various characteristic information of the navigable bridge area can be obtained, including environmental characteristics, bridge site characteristics, and ship characteristics. These multi-source data are integrated to form historical characteristic data; the historical characteristic data is divided into two parts:

[0052] Single object features: Cover the static and dynamic features of a single ship, as well as features such as the span of a navigable bridge that do not consider the coupling relationship between similar objects; Based on these single object features, a trajectory prediction model for a single ship is constructed through a graph neural network.

[0053] Traffic flow features: Used to consider the coupling relationship between ships; Given the complexity of ship traffic flow features, a multi-level Monte Carlo sampling method is used to generate random ship flow simulation results, which are incorporated into the graph neural network as new features and combined with the single ship trajectory prediction model to form a ship cluster trajectory prediction model;

[0054] S22: Model prediction and training: As Figure 3 shown, the data collected by the monitoring field at time t is used as features and input into the model. By continuously looping, the prediction of the subsequent trajectory of the ship can be achieved; At the same time, the predicted data can be used as new input for model training to achieve a data-driven effect.

[0055] In the specific implementation, in S1, MMSI, as the internationally unified unique ship code, can be defined as the label of the subsequent target detection model. This module identifies the basic information of the ship and serves as the label in the target instance segmentation and target tracking process for facilitating subsequent data integration.

[0056] In the specific implementation, in S1, the information collected by the environmental information collection module also includes bridge and bridge site information, including bridge navigation hole information and shoreline projects.

[0057] In the specific implementation, in S21, when integrating multi-source data, data fusion technologies, including Kalman filtering and Bayesian fusion, are used to fuse data from different sources to improve the accuracy and reliability of the data.

[0058] In the specific implementation, in S21, feature selection and dimensionality reduction processing are performed on historical feature data to remove redundant features and retain features that have an important impact on trajectory prediction. The principal component analysis PCA and linear discriminant analysis LDA methods are used for feature selection and dimensionality reduction.

[0059] In the specific implementation, in S21, an attention mechanism is introduced into the graph neural network to enable the model to pay more attention to features and nodes that are important for trajectory prediction. The attention mechanism can adaptively allocate weights to different features and nodes, improving the prediction accuracy of the model.

[0060] In the specific implementation manner, in S21, a reinforcement learning algorithm is used to optimize the model. Reinforcement learning can continuously adjust the parameters of the model through interaction with the environment to achieve optimal trajectory prediction. The navigation trajectory of the ship is regarded as the state, and the prediction accuracy is regarded as the reward, and the model is continuously optimized through the reinforcement learning algorithm.

[0061] In the specific implementation manner, in S22, in addition to using the prediction data for model training, the prediction results can also be compared and analyzed with the actual navigation trajectory to evaluate the prediction accuracy of the model. The mean square error (MSE) and mean absolute error (MAE) metrics are used to evaluate the performance of the model.

[0062] In the specific implementation manner, in S21, the data integration of the collected information is achieved through the following methods:

[0063] S211: Define the data sources and types;

[0064] S212: Associate with the ship as the core. Using the MMSI code of the ship as the key identifier, associate the basic ship information obtained by the AIS data receiver with the ship appearance and trajectory information obtained through object instance segmentation and object tracking in the image information acquisition module, so that different types of data of the same ship can be corresponding; for radar data, it can also be matched and associated with the ships in the AIS and image information according to information such as the position of the ship. Especially in low visibility conditions, use radar data to supplement and correct the deficiencies of the ship position information in other data.

[0065] S213: Integrate the environmental information and ship information. Integrate the water level, flow rate, wind speed, and bridge and bridge site information obtained by the environmental information acquisition module with the relevant information of the ship; combine the real-time water level and flow rate information with the navigation dynamic information of the ship to analyze the influence of the water flow on the ship's navigation trajectory; associate the wind speed information with the characteristics such as the stability of the ship and consider the effect of the wind on the ship's navigation.

[0066] S214: Construct historical feature data;

[0067] S215: Data preprocessing and normalization.

[0068] In the specific implementation manner, in S22, the specific process of constructing the model is as follows:

[0069] S221: Determine the nodes and edges of the graph;

[0070] Ship nodes: Each ship is regarded as a node in the graph. The node features can include the static information of the ship, such as ship type, size, load, etc.; dynamic information, such as position, speed, heading, etc.; and voyage information and safety information in the AIS data.

[0071] Environmental nodes: Consider environmental factors as nodes. For example, the locations of water level gauges and wind speed sensors can be regarded as environmental nodes. Their features include the real-time water level and flow velocity measured by the water level gauge, and the real-time wind speed obtained by the wind speed sensor, etc.

[0072] Bridge and bridge site nodes: Set the positions of the navigation holes of the bridge, shoreline projects, etc. as nodes. The node features include information such as the bridge span, the size of the navigation hole, and the shape of the shoreline;

[0073] Edges between ships: Determine the connection of the edges according to the relative positions, distances, etc. between ships. For example, when the distance between two ships is within a certain range, it is considered that there is an edge between them. The features of the edge can be the distance, relative speed, included angle, etc. between the two ships.

[0074] Edges between ships and the environment: If a ship is within the influence range of a water level gauge or a wind speed sensor, an edge is established between the ship node and the corresponding environmental node. The features of the edge can be the distance between the ship and the environmental monitoring point, the degree of influence of environmental factors on the ship, etc.

[0075] Edges between ships and bridges and bridge sites: When a ship approaches the navigation hole of a bridge or a shoreline project, an edge is established between the ship node and the corresponding bridge and bridge site nodes. The features of the edge can include the distance between the ship and the navigation hole or the shoreline, the included angle between the ship's heading and the axis of the navigation hole, etc.

[0076] S222: Select the graph neural network architecture: Select the GCN graph convolutional neural network or the GAT graph attention network or GraphSAGE;

[0077] S223: Design the layer structure of the model:

[0078] Input layer: Input the initial feature vectors of the nodes into the model, including various attribute features of the ship, environmental parameters, and bridge and bridge site information. These features undergo preliminary linear transformation or embedding mapping through the input layer and are converted into low-dimensional vector representations suitable for graph neural network processing;

[0079] Hidden layer: Stack multiple graph neural network layers, such as GCN layers, GAT layers, or GraphSAGE layers, as hidden layers; In each layer, by aggregating the features of neighbor nodes and updating the features of the current node, gradually extract the high-order relationships and complex patterns in the graph data; The number of hidden layers and the number of neurons need to be adjusted according to the specific problem and data characteristics, and generally, the optimal structure is determined through experiments;

[0080] Output layer: Design the output layer according to the objectives of the prediction task; if predicting the trajectory of a ship, the output layer can be a fully connected layer that outputs the position coordinates, speed, and heading information of the ship at a future time step; if predicting events such as whether a ship will collide, the output layer can be a binary classification layer that outputs a probability value indicating the likelihood of a collision.

[0081] S224: Define the loss function and optimizer;

[0082] Loss function

[0083] Regression loss: If the output of the model is continuous trajectory data, etc., the mean squared error (MSE) loss function can be used to measure the difference between the predicted value and the true value. For example, calculate the mean squared error between the predicted ship position and the actual observed position.

[0084] Classification loss: For classification tasks, such as predicting whether a ship can safely pass through a navigation bridge area, etc., the cross-entropy loss function can be adopted to measure the difference between the predicted probability distribution and the true label.

[0085] Optimizer: Select a suitable optimizer to update the model's parameters, such as Stochastic Gradient Descent (SGD) and its variants Adagrad, Adadelta, RMSProp, Adam, etc. The Adam optimizer usually performs well in many scenarios. It can adaptively adjust the learning rate and update the parameters according to the gradient history information of the parameters, which helps the model converge faster.

[0086] S225: Model training and adjustment.

[0087] Data partitioning: Divide the collected multi-source data into a training set, a validation set, and a test set. The training set is used to train the model and let the model learn the patterns and rules in the data; the validation set is used to adjust the hyperparameters of the model during training, such as selecting the optimal number of hidden layers, learning rate, etc.; the test set is used to evaluate the performance of the model on unseen data to ensure that the model has good generalization ability.

[0088] Model training: Use the training set to train the constructed graph neural network model. Calculate the gradient of the loss function with respect to the model parameters through the backpropagation algorithm and use the optimizer to update the parameters. Continuously iterate the training to gradually reduce the loss function of the model and continuously improve the performance of the model.

[0089] Hyperparameter tuning: According to the performance on the validation set, adjust the hyperparameters of the model, such as trying different graph neural network architectures, the number of hidden layers, the number of neurons, the learning rate, the number of attention heads, etc. Through multiple experiments and comparisons, find the combination of hyperparameters that makes the model performance optimal.

[0090] Model evaluation: Evaluate the trained model on the test set, and use evaluation metrics such as accuracy, recall, F1 score, mean squared error, etc. to measure the performance of the model, ensuring that the model can accurately predict the trajectories of ships in the navigable bridge area and meet the requirements of practical applications.

[0091] The components not elaborated in this embodiment are all existing components that can be purchased through public channels.

[0092] The above description of the embodiments is for the convenience of those of ordinary skill in the art to understand and use the invention. It is obvious that those skilled in the art can easily make various modifications to these embodiments and apply the general principles described herein to other embodiments without creative efforts. Therefore, the present invention is not limited to the above embodiments, and the improvements and modifications made by those skilled in the art without departing from the scope of the present invention according to the disclosure of the present invention should be within the protection scope of the present invention.

Claims

1. A method for predicting ship trajectories in a navigable bridge area based on graph neural network, characterized in that: The following steps are involved: S1: Set up a ship monitoring field in the navigation bridge area to obtain ship information and scene information; the monitoring field includes: image information acquisition module, ship basic information acquisition module, and environmental information acquisition module; The image information acquisition module includes a bridge deck main camera and a shoreline auxiliary camera. The image information acquisition module uses target instance segmentation and target tracking to obtain information; The ship basic information collection module includes an AIS data receiver and radar; the static information, dynamic information, voyage information and safety information of the ship are obtained through AIS; the radar is used to make up for the defect of low recognition accuracy of the image information collection module under low visibility conditions; The environmental information collection module includes a water level meter and a wind speed sensor. The water level meter is used to obtain the real-time water level and flow rate, and the wind speed sensor is used to obtain the real-time wind speed; S2: Establish a ship prediction model in the navigable bridge area based on graph neural network: S21: Data integration and model building: The entire monitoring system can be used to obtain various characteristic information of the navigable bridge area, including environmental characteristics, bridge site characteristics and ship characteristics. These multi-source data are integrated to form historical characteristic data; the historical characteristic data is divided into two parts: Single object features: covers the static and dynamic features of a single ship, as well as the span of a navigable bridge without considering the coupling relationship between similar objects; based on these single object features, a trajectory prediction model for a single ship is constructed through a graph neural network; Traffic flow characteristics: used to consider the coupling relationship between ships. In view of the complexity of ship traffic flow characteristics, a multi-level Monte Carlo sampling method is used to generate random ship flow simulation results, which are incorporated into the graph neural network as new features and combined with a single ship trajectory prediction model to form a ship cluster trajectory prediction model. S22: Model prediction and training: The data collected by the monitoring field at time t is input into the model as features. Through continuous cycles, the subsequent trajectory of the ship can be predicted. At the same time, the predicted data can be used as new input for model training to achieve a data-driven effect.

2. According to claim 1, a method for predicting ship trajectories in a navigable bridge area based on a graph neural network is characterized in that: In S1, MMSI, as an internationally unified ship unique code, can be defined as the label of the subsequent target detection model. This module identifies the basic information of the ship and serves as a label in the target instance segmentation and target tracking process to facilitate subsequent data integration.

3. The method for predicting ship trajectories in a navigable bridge area based on graph neural network according to claim 1 is characterized in that: In S1, the information collected by the environmental information collection module also includes bridge and bridge site information, including bridge navigation hole information, and shoreline engineering.

4. According to the method for predicting ship trajectories in a navigable bridge area based on graph neural network in claim 1, it is characterized in that: In S21, when integrating multi-source data, data fusion technology, including Kalman filtering and Bayesian fusion, is used to fuse data from different sources to improve the accuracy and reliability of the data.

5. The method for predicting ship trajectories in a navigable bridge area based on graph neural network according to claim 1 is characterized in that: In S21, feature selection and dimensionality reduction are performed on the historical feature data to remove redundant features and retain features that have an important impact on trajectory prediction. Principal component analysis (PCA) and linear discriminant analysis (LDA) methods are used for feature selection and dimensionality reduction.

6. The method for predicting ship trajectories in a navigable bridge area based on graph neural network according to claim 1 is characterized in that: In S21, the attention mechanism is introduced into the graph neural network, so that the model can pay more attention to the features and nodes that are important for trajectory prediction.

7. The method for predicting ship trajectories in a navigable bridge area based on graph neural network according to claim 1 is characterized in that: In S21, a reinforcement learning algorithm is used to optimize the model. Reinforcement learning can continuously adjust the parameters of the model through interaction with the environment to achieve the optimal trajectory prediction. The ship's navigation trajectory is used as the state, and the accuracy of the prediction is used as the reward, and the model is continuously optimized through the reinforcement learning algorithm.

8. The method for predicting ship trajectories in a navigable bridge area based on graph neural network according to claim 1 is characterized in that: In S22, in addition to using the predicted data for model training, the predicted results can also be compared and analyzed with the actual navigation trajectory to evaluate the prediction accuracy of the model. The mean square error (MSE) and mean absolute error (MAE) indicators are used to evaluate the performance of the model.

9. The method for predicting ship trajectories in a navigable bridge area based on graph neural network according to claim 1 is characterized in that: In S21, data integration of the collected information is achieved by: S211: Clarify the data source and type; S212: With the ship as the core, the MMSI code of the ship is used as the key identifier to associate the basic information of the ship obtained by the AIS data receiver with the appearance and trajectory information of the ship obtained by target instance segmentation and target tracking in the image information acquisition module, so that different types of data of the same ship can be matched; for radar data, the ship's position information can also be matched and associated with the ship in the AIS and image information, especially when visibility is low, the radar data can be used to supplement and correct the lack of ship position information in other data; S213: Integrate environmental information with ship information, and integrate the water level, flow velocity, wind speed, bridge and bridge site information obtained by the environmental information acquisition module with the relevant information of the ship; combine the real-time water level and flow velocity information with the ship's navigation dynamics information to analyze the impact of water flow on the ship's navigation trajectory; associate the wind speed information with the stability characteristics of the ship to consider the effect of wind on the ship's navigation. S214: construct historical feature data; S215: Data preprocessing and normalization.

10. The method for predicting ship trajectories in a navigable bridge area based on graph neural network according to claim 1, characterized in that: In S22, the specific process of building the model is: S221: Determine the nodes and edges of the graph; S222: Select graph neural network architecture: select GCN graph convolutional neural network or GAT graph attention network or GraphSAGE; S223: Design model layer structure: Input layer: The initial feature vector of the node is input into the model, and various attribute characteristics of the ship, environmental parameters, bridge and bridge site information are initially linearly transformed or embedded in the input layer to convert them into low-dimensional vector representations suitable for graph neural network processing; Hidden layer: stack multiple graph neural network layers, with GCN layers, GAT layers or GraphSAGE layers as hidden layers; in each layer, by aggregating the features of neighboring nodes and updating the features of the current node, high-order relationships and complex patterns in the graph data are gradually extracted; the number of hidden layers and the number of neurons need to be adjusted according to the specific problem and data characteristics, and the optimal structure is generally determined through experiments; Output layer: The output layer is designed according to the goal of the prediction task. If the goal is to predict the trajectory of a ship, the output layer can be a fully connected layer that outputs the position coordinates, speed, and heading information of the ship at a certain time step in the future. If the goal is to predict whether a ship will collide, the output layer can be a binary classification layer that outputs a probability value indicating the possibility of a collision. S224: define loss function and optimizer; S225: Model training and tuning.

Images