Ship navigation behavior identification method and system, electronic equipment and storage medium
By fusion and screening of ship position data from multiple signal sources, combined with normal ship trajectory model, the ship's navigation behavior is identified, data conflicts and behavior recognition problems are solved, and more accurate navigation trajectory generation and abnormal detection are achieved.
Patent Information
- Application Number
- CN202411387234.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-30
- Publication Date
- 2024-11-15
AI Technical Summary
The prior art is difficult to effectively integrate ship position data provided by multiple signal sources, resulting in data conflicts and analysis difficulties, and it is difficult to accurately identify the normal or abnormal navigation behavior of the ship.
By acquiring ship position data from multiple signal sources (AIS, radar, satellite navigation systems) for data fusion, filtering the optimal data to generate navigation trajectories, and using the pre-constructed normal ship trajectory model to calculate the trajectory feature similarity and identify navigation behavior.
It improves positioning accuracy and data integrity, the generated navigation trajectory is more accurate, and it can timely identify abnormal navigation behaviors, improving safety and monitoring efficiency.
Smart Images

Figure CN118965235A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of ship monitoring technology or other related fields, and in particular to a method and system for identifying ship navigation behavior, an electronic device and a storage medium. Background Art
[0002] In order to ensure maritime traffic safety and optimize route planning, it is crucial to obtain and process the position data of target ships to monitor their navigation behavior. Multi-source signal technologies such as the Automatic Identification System (AIS), radar systems, and satellite navigation systems provide a reliable means for obtaining the position of ships in real time. These systems each have their own advantages and limitations: AIS provides rich information such as ship identity, speed, and heading, but its coverage is limited by shore-based base stations or satellite reception; radar systems can effectively detect within the line of sight, but are affected by weather conditions and do not provide identity information; satellite navigation systems can provide high-precision positioning worldwide, but rely on the integrity of ship equipment and reception conditions.
[0003] However, the existing technology has the following technical problems: the ship position data provided by different signal sources have different formats, accuracy, and update frequency, and direct use may lead to data conflicts and analysis difficulties; how to effectively integrate the ship position data from multiple signal sources, eliminate redundancy and contradictions, and generate a navigation trajectory that is closest to the actual situation is a major challenge faced by the existing technology; the boundary between normal navigation behavior and abnormal navigation behavior is blurred, and it is necessary to build a reasonable ship trajectory model based on a large amount of historical data and professional knowledge, and design effective feature extraction and comparison methods to identify abnormal navigation behavior and conduct timely alarm processing.
[0004] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention
[0005] The embodiments of the present invention provide a method and system for identifying the navigation behavior of a ship, an electronic device and a storage medium, so as to at least solve the technical problem that the navigation behavior of a ship cannot be accurately identified in the related art.
[0006] According to one aspect of an embodiment of the present invention, a method for identifying ship navigation behavior is provided, comprising: acquiring position data of a target ship collected by N signal sources, the signal sources comprising: a ship automatic identification system, a radar system and a satellite navigation system, and N is a positive integer greater than 1; fusing the position data of the target ship under different signal sources to obtain fused position data; in each spatiotemporal region where the target ship is sailing, screening out optimal position data from the fused position data, and generating a navigation track of the target ship based on the optimal position data; inputting the navigation track of the target ship into a pre-constructed normal ship track model, and calculating the trajectory feature similarity between the navigation track of the target ship and the navigation track of the normal ship, so as to identify the normal navigation behavior or abnormal navigation behavior of the target ship.
[0007] Optionally, the position data of the target ship under different signal sources are fused to obtain fused position data, including: preprocessing the position data of the target ship under different signal sources to obtain preprocessed position data, wherein the data preprocessing includes at least one of the following: removing invalid data, removing duplicate data, removing erroneous data, and converting the position data of different signal sources into a unified data format, and time aligning the position data of different signal sources; using a data fusion algorithm to perform fusion calculation on the preprocessed position data to obtain the fused position data.
[0008] Optionally, a data fusion algorithm is used to perform fusion calculation on the preprocessed ship position data to obtain the fused ship position data, including: assigning corresponding weight parameters to the ship position data of different signal sources according to the accuracy level, update frequency and historical reliability of different signal sources, performing weighted average calculation according to the weights of the ship position data of different signal sources in the same time window, and using the calculated weighted average data as the fused ship position data; and / or, using a Kalman filter to recursively update the ship position data of different signal sources, and using the output optimal estimation data as the fused ship position data; and / or, using a particle filtering algorithm to simulate a particle swarm to characterize the position state of the ship, updating the particle weight distribution according to the ship position data of different signal sources, and using the particle data with the largest weight as the fused ship position data.
[0009] Optionally, in each spatiotemporal region where the target ship is sailing, the step of selecting the optimal ship position data from the fused ship position data, and generating the navigation track of the target ship based on the optimal ship position data comprises: dividing the time axis during the navigation of the target ship into a plurality of continuous time periods, each time period corresponding to a spatiotemporal region, wherein the spatiotemporal region represents the maximum geographical range within which the target ship can move within the corresponding time period; in each of the spatiotemporal regions, selecting one of the ship position data from the fused ship position data based on a predetermined reference standard as the optimal ship position data for the spatiotemporal region, wherein the predetermined reference standard comprises at least one of the following: selecting ship position data whose accuracy level, update frequency, and historical reliability index are greater than an index threshold during the data fusion process, selecting ship position data with a smooth transition between adjacent spatiotemporal regions and a continuity greater than a predetermined continuity threshold, and selecting ship position data that is consistent with the expected navigation mode of the target ship; and connecting the optimal ship position data in each of the spatiotemporal regions in chronological order to generate the navigation track of the target ship.
[0010] Optionally, the navigation trajectory of the target ship is input into a pre-constructed normal ship trajectory model, and the trajectory feature similarity between the navigation trajectory of the target ship and the navigation trajectory of the normal ship is calculated to identify the normal navigation behavior or abnormal navigation behavior of the target ship, including: converting the navigation trajectory of the target ship into a feature representation consistent with the normal ship trajectory model, and using a distance measurement function or a machine learning model to calculate the trajectory feature similarity between the navigation trajectory characteristics of the target ship and the trajectory model of the normal ship; comparing the trajectory feature similarity with a preset similarity threshold; if the trajectory feature similarity is greater than or equal to the similarity threshold, it is determined that the current navigation state of the target ship is in normal navigation behavior; if the trajectory feature similarity is less than the similarity threshold, it is determined that the current navigation state of the target ship has abnormal navigation behavior.
[0011] Optionally, the method for identifying ship navigation behavior also includes: obtaining performance indicators of different signal sources, wherein the performance indicators include at least one of the following: signal strength, data update frequency, and noise level; inputting the performance indicators of different signal sources into a pre-constructed weight prediction model to obtain real-time weights of different signal sources, wherein the weight prediction model is trained using a neural network or a machine learning algorithm using multiple first training data, and the first training data is derived from performance data of historical signal sources and weights set by expert experience; according to the real-time weights of different signal sources, data fusion is performed on the ship position data within the same time window to obtain fused ship position data.
[0012] Optionally, the method for identifying ship navigation behavior also includes: obtaining multi-source information of the target ship, wherein the multi-source information includes the position data, ship communication data, navigation log and meteorological data of the target ship; inputting the multi-source information of the target ship into a trained ship navigation behavior recognition model to obtain the current navigation behavior category and probability of the target ship, wherein the ship navigation behavior recognition model is trained by a deep learning model using multiple second training data, and the second training data includes multi-source information and corresponding known navigation behavior labels.
[0013] According to another aspect of an embodiment of the present invention, a system for identifying ship navigation behavior is also provided, including: a position data acquisition unit, used to acquire the position data of a target ship collected by N signal sources, the signal sources including: a ship automatic identification system, a radar system and a satellite navigation system, N being a positive integer greater than 1; a data fusion unit, used to fuse the position data of the target ship under different signal sources to obtain fused position data; a navigation trajectory generation unit, used to filter out optimal position data from the fused position data in each spatiotemporal area where the target ship navigates, and generate the navigation trajectory of the target ship based on the optimal position data; a navigation behavior discrimination module, used to input the navigation trajectory of the target ship into a pre-constructed normal ship trajectory model, calculate the trajectory feature similarity between the navigation trajectory of the target ship and the navigation trajectory of the normal ship, so as to identify the normal navigation behavior or abnormal navigation behavior of the target ship.
[0014] Optionally, the data fusion unit includes: a preprocessing module, used to perform data preprocessing on the position data of the target ship under different signal sources to obtain preprocessed position data, wherein the data preprocessing includes at least one of the following: removing invalid data, removing duplicate data, removing erroneous data, and converting the position data of different signal sources into a unified data format, and time aligning the position data of different signal sources; a data fusion module, used to use a data fusion algorithm to perform fusion calculation on the preprocessed position data to obtain the fused position data.
[0015] Optionally, the data fusion module includes: a weight parameter assignment submodule, which is used to assign corresponding weight parameters to the ship position data of different signal sources according to the accuracy level, update frequency and historical reliability of different signal sources, and for the ship position data of different signal sources in the same time window, weighted average calculation is performed according to their weights, and the calculated weighted average data is used as the fused ship position data; and / or, a ship position data update submodule, which is used to recursively update the ship position data of different signal sources using a Kalman filter, and use the output optimal estimation data as the fused ship position data; and / or, a ship position fusion submodule, which is used to adopt a particle filtering algorithm, simulate a particle swarm to characterize the position state of the ship, update the particle weight distribution according to the ship position data of different signal sources, and use the particle data with the largest weight as the fused ship position data.
[0016] Optionally, the navigation trajectory generating unit comprises: a time axis division module, used for dividing the time axis during the navigation of the target ship into a plurality of continuous time periods, each time period corresponding to a space-time region, wherein the space-time region represents the maximum geographical range within which the target ship can move within the corresponding time period; a ship position data selection module, used for selecting one of the ship position data from the fused ship position data in each of the space-time regions based on a predetermined reference standard as the optimal ship position data for the space-time region, wherein the predetermined reference standard comprises at least one of the following: selecting ship position data whose accuracy level, update frequency, and historical reliability index in the data fusion process are greater than an index threshold, selecting ship position data with a smooth transition between adjacent space-time regions and a continuity greater than a predetermined continuity threshold, and selecting ship position data that is consistent with the expected navigation mode of the target ship; a navigation trajectory generating module, used for connecting the optimal ship position data in each of the space-time regions in chronological order to generate the navigation trajectory of the target ship.
[0017] Optionally, the navigation behavior discrimination module includes: a trajectory feature similarity calculation submodule, used to convert the navigation trajectory of the target ship into a feature representation consistent with the normal ship trajectory model, and use a distance measurement function or a machine learning model to calculate the trajectory feature similarity between the navigation trajectory characteristics of the target ship and the trajectory model of the normal ship; a similarity comparison submodule, used to compare the trajectory feature similarity with a preset similarity threshold; a first navigation behavior discrimination submodule, used to determine that the current navigation state of the target ship is in normal navigation behavior when the trajectory feature similarity is greater than or equal to the similarity threshold; a second navigation behavior discrimination submodule, used to determine that the current navigation state of the target ship has abnormal navigation behavior when the trajectory feature similarity is less than the similarity threshold.
[0018] Optionally, the ship navigation behavior identification system also includes: a performance indicator acquisition unit, used to obtain performance indicators of different signal sources, wherein the performance indicators include at least one of the following: signal strength, data update frequency, noise level; a signal source weight determination unit, used to input the performance indicators of different signal sources into a pre-constructed weight prediction model to obtain real-time weights of different signal sources, wherein the weight prediction model is obtained by training with a neural network or a machine learning algorithm using multiple first training data, and the first training data is derived from the performance data of historical signal sources and the weights set by expert experience; a ship position data fusion unit, used to perform data fusion on the ship position data within the same time window according to the real-time weights of different signal sources to obtain fused ship position data.
[0019] Optionally, the ship navigation behavior identification system also includes: a multi-source information acquisition unit, used to acquire multi-source information of the target ship, wherein the multi-source information includes the position data, ship communication data, navigation log and meteorological data of the target ship; a navigation behavior identification unit, used to input the multi-source information of the target ship into a trained ship navigation behavior identification model to obtain the current navigation behavior category and probability of the target ship, wherein the ship navigation behavior identification model is trained by a deep learning model using multiple second training data, and the second training data includes multi-source information and corresponding known navigation behavior labels.
[0020] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is further provided, wherein the computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the above-mentioned methods for identifying ship navigation behaviors.
[0021] According to another aspect of an embodiment of the present invention, there is also provided an electronic device, comprising one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement any one of the above-mentioned methods for identifying ship navigation behaviors.
[0022] According to another aspect of an embodiment of the present invention, there is further provided a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the steps of any one of the above-mentioned methods for identifying the navigation behavior of a ship are implemented.
[0023] In the present disclosure, the position data of a target ship collected by N signal sources are obtained, the signal sources include: a ship automatic identification system, a radar system and a satellite navigation system, and N is a positive integer greater than 1; the position data of the target ship under different signal sources are fused to obtain fused position data; in each spatiotemporal region where the target ship is sailing, the optimal position data is screened out from the fused position data, and the navigation track of the target ship is generated based on the optimal position data; the navigation track of the target ship is input into a pre-constructed normal ship trajectory model, and the trajectory feature similarity between the navigation track of the target ship and the navigation track of the normal ship is calculated to identify the normal navigation behavior or abnormal navigation behavior of the target ship.
[0024] Through the above steps, by integrating the ship position data from different signal sources such as AIS, radar, and navigation systems, the positioning accuracy and data integrity are improved, providing a more accurate basis for the subsequent generation of navigation trajectories; in each spatiotemporal area where the target ship is sailing, by setting reasonable screening rules, the optimal data points are selected from the fused ship position data to construct a continuous, smooth navigation trajectory that represents the actual navigation path, thereby overcoming the local distortion or discontinuity that may be caused by a single data source, and the generated trajectory is more in line with the actual movement law of the ship.
[0025] In the present disclosure, a pre-constructed normal ship trajectory model can be used to calculate the trajectory feature similarity between the target ship's navigation trajectory and the typical normal trajectory in the model; by setting a threshold or using a classification algorithm, it is possible to determine whether the navigation behavior of the target ship is within the normal range, thereby improving the sensitivity and accuracy of abnormal navigation behavior detection, helping to promptly discover potential dangerous operations, violations or security threats, and thus solving the technical problem that related technologies cannot accurately identify ship navigation behaviors. BRIEF DESCRIPTION OF THE DRAWINGS
[0026] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:
[0027] Figure 1 is a flow chart of an optional method for identifying ship navigation behavior according to an embodiment of the present invention;
[0028] Figure 2 is a schematic diagram of an optional ship navigation behavior recognition system according to an embodiment of the present invention;
[0029] Figure 3 The present invention is a hardware structure block diagram of an electronic device (or mobile device) for a method for identifying ship navigation behavior according to an embodiment of the present invention. DETAILED DESCRIPTION
[0030] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.
[0031] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0032] It should be noted that the information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) collected by this disclosure are information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of relevant data are in compliance with the relevant laws, regulations and standards of the relevant regions, necessary confidentiality measures are taken, and public order and good customs are not violated, and corresponding operation portals are provided for users to choose to authorize or refuse. For example, an interface is set up between this system and relevant users or organizations. Before obtaining relevant information, it is necessary to send an acquisition request to the aforementioned user or organization through the interface, and obtain relevant information after receiving the consent information fed back by the aforementioned user or organization.
[0033] It should be noted that in the present disclosure, when collecting and analyzing customer information, corresponding operation entrances are provided for users to choose to agree or reject the automated decision results; if the user chooses to reject, the expert decision-making process is entered.
[0034] The following embodiments of the present invention can be applied to various ship navigation behavior identification systems / applications / devices. The present invention improves positioning accuracy and data integrity by integrating ship position data from different signal sources such as AIS, radar, and navigation systems, providing a more accurate basis for subsequent navigation trajectory generation; at the same time, in each time and space area where the target ship is sailing, by setting reasonable screening rules, the optimal data points are selected from the fused ship position data to construct a continuous, smooth navigation trajectory that represents the actual navigation path, thereby overcoming the local distortion or discontinuity that may be caused by a single data source, and the generated trajectory is more in line with the actual movement law of the ship.
[0035] At the same time, the present invention can use a pre-constructed normal ship trajectory model to calculate the trajectory feature similarity between the target ship's navigation trajectory and the typical normal trajectory in the model. By setting a threshold or using a classification algorithm, it can be judged whether the navigation behavior of the target ship is within the normal range, thereby improving the sensitivity and accuracy of abnormal navigation behavior detection, which helps to timely discover potential dangerous operations, violations or security threats.
[0036] The present invention is described in detail below in conjunction with various embodiments.
[0037] Embodiment 1
[0038] According to an embodiment of the present invention, an embodiment of a method for identifying ship navigation behavior is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.
[0039] Figure 1 is a flow chart of an optional method for identifying ship navigation behavior according to an embodiment of the present invention, such as Figure 1 As shown, the method comprises the following steps:
[0040] Step S101, obtaining the position data of the target ship collected by N signal sources, the signal sources include: a ship automatic identification system, a radar system and a satellite navigation system, and N is a positive integer greater than 1.
[0041] Step S102, fusing the target ship's position data from different signal sources to obtain fused position data.
[0042] Step S103, in each spatiotemporal area where the target ship is sailing, the optimal ship position data is screened out from the fused ship position data, and the navigation track of the target ship is generated based on the optimal ship position data.
[0043] Step S104, inputting the navigation track of the target ship into a pre-built normal ship trajectory model, calculating the trajectory feature similarity between the navigation track of the target ship and the navigation track of the normal ship, so as to identify the normal navigation behavior or abnormal navigation behavior of the target ship.
[0044] Through the above steps, the position data of the target ship collected by N signal sources can be obtained, and the signal sources include: a ship automatic identification system, a radar system, and a satellite navigation system. The position data of the target ship under different signal sources are fused to obtain the fused position data. In each time and space region where the target ship is sailing, the optimal position data is screened out from the fused position data, and the navigation track of the target ship is generated based on the optimal position data. The navigation track of the target ship is input into a pre-constructed normal ship trajectory model, and the similarity of the trajectory features between the navigation track of the target ship and the navigation track of the normal ship is calculated to identify the normal navigation behavior or abnormal navigation behavior of the target ship. In this embodiment, the pre-constructed normal ship trajectory model is used to calculate the similarity of the trajectory features between the navigation track of the target ship and the typical normal trajectory in the model. By setting a threshold or using a classification algorithm, it can be judged whether the navigation behavior of the target ship belongs to the normal range, thereby improving the sensitivity and accuracy of abnormal navigation behavior detection, helping to timely discover potential dangerous operations, violations or security threats, and thus solving the technical problem that the navigation behavior of the ship cannot be accurately identified in the related technology.
[0045] The embodiment of the present invention is described in detail below in combination with the above steps.
[0046] Step S101, obtaining the position data of the target ship collected by N signal sources, the signal sources including: a ship automatic identification system, a radar system and a satellite navigation system.
[0047] It should be noted that the satellite navigation system mentioned in this embodiment can be selected by oneself, for example, the Beidou navigation system and the GPS system.
[0048] The above-mentioned position data refers to the accurate information used to describe the geographical location of a ship active at sea on the surface of the earth at a specific moment or period. The position data includes but is not limited to the geographical location, timestamp, accuracy information, data source identification and auxiliary information. For example, geographical location: the core content of the position data is the longitude and latitude coordinates of the ship, which accurately calibrates the two-dimensional position of the ship on the earth. The longitude and latitude coordinates are in degrees, usually including longitude (east-west direction) and latitude (north-south direction), and may also have altitude or depth information (for submersible ships or underwater equipment), which together constitute the three-dimensional positioning of the ship on the surface of the earth. Timestamp: The position data must be associated with a specific time point or time period in order to accurately record the moment when the ship's position occurs. The timestamp is usually expressed in international standard time (such as Coordinated Universal Time, UTC) and is accurate to seconds or even milliseconds to ensure the real-time and traceability of the data. Accuracy information: The position data may be accompanied by a position accuracy indicator, which reflects the uncertainty of measuring or calculating the position. It is usually given in the form of error radius (such as circular error probability, CEP) or standard deviation of longitude and latitude to indicate the credibility of the data and the performance of the positioning system. Data source identification: For ship position data from different signal sources (such as ship automatic identification system AI S, radar, satellite navigation system such as Beidou / GPS), the data source identification may be included to distinguish the data source, which is helpful for subsequent data fusion, analysis and troubleshooting. Auxiliary information: In addition to basic location information, ship position data may also contain other auxiliary information related to the ship's status, such as speed, heading, navigation mode (such as anchoring, sailing, anchoring, etc.), ship type, size, destination, estimated arrival time, etc., which helps to fully understand the ship's navigation dynamics and intentions.
[0049] Step S102, fusing the target ship's position data from different signal sources to obtain fused position data.
[0050] Optionally, step S102 includes: performing data preprocessing on the position data of the target ship under different signal sources to obtain preprocessed position data, wherein the data preprocessing includes at least one of the following: removing invalid data, removing duplicate data, removing erroneous data, and converting the position data of different signal sources into a unified data format, and time aligning the position data of different signal sources; using a data fusion algorithm to perform fusion calculation on the preprocessed position data to obtain fused position data.
[0051] The data preprocessing mentioned in this embodiment includes but is not limited to removing invalid, repeated or obviously erroneous ship position data points; converting the ship position data from different signal sources into a unified data format, including timestamp, longitude and latitude, altitude (if any), speed, heading, etc.; and time aligning the ship position data from different sources to ensure fusion within the same time window.
[0052] It should be noted that when using the data fusion algorithm to perform fusion calculation on the pre-processed ship position data to obtain the fused ship position data, it includes: assigning corresponding weight parameters to the ship position data of different signal sources according to the accuracy level, update frequency and historical reliability of different signal sources, and performing weighted average calculation according to the weights of the ship position data of different signal sources in the same time window, and using the calculated weighted average data as the fused ship position data; and / or, using the Kalman filter to recursively update the ship position data of different signal sources, and using the output optimal estimation data as the fused ship position data; and / or, using the particle filtering algorithm to simulate the particle swarm to characterize the position state of the ship, updating the particle weight distribution according to the ship position data of different signal sources, and using the particle data with the largest weight as the fused ship position data.
[0053] In this embodiment, the ship position data from different signal sources can be cleaned and formatted to ensure that the timestamps of each data set are consistent or comparable, the position data accuracy is adapted, and obvious errors or outliers are removed. Use suitable multi-source ship position data fusion algorithms, such as Kalman filtering, particle filtering, Bayesian estimation, etc. These algorithms can combine the characteristics and uncertainties of each source data to generate fused ship position data and improve positioning accuracy and reliability. Take AIS, radar and navigation data as input, run the selected fusion algorithm, and output the fused ship position data set, including the optimal position estimate at each time point.
[0054] Optionally, a weight w is assigned to each signal source based on its accuracy level, update frequency, historical reliability, etc. i Signal sources with high accuracy, high update frequency, and good historical reliability should be given greater weights. Make sure the sum of all weights equals 1.
[0055] For each ship position data p provided by the signal source i in the same time window i (including longitude and latitude information), according to its weight w i Perform weighted average calculation:
[0056]
[0057] Where N represents the number of signal sources involved in the fusion.
[0058] Step S103, in each spatiotemporal area where the target ship is sailing, the optimal ship position data is screened out from the fused ship position data, and the navigation track of the target ship is generated based on the optimal ship position data.
[0059] Optionally, step S103 includes: dividing the time axis during the navigation of the target ship into multiple continuous time periods, each time period corresponding to a space-time area, wherein the space-time area represents the maximum geographical range within which the target ship can move within the corresponding time period; in each space-time area, selecting one of the ship position data from the fused ship position data based on a predetermined reference standard as the optimal ship position data for the space-time area, wherein the predetermined reference standard includes at least one of the following: selecting ship position data whose accuracy level, update frequency, and historical reliability index during the data fusion process are greater than an index threshold, selecting ship position data with a smooth transition between adjacent space-time areas and a continuity greater than a predetermined continuity threshold, and selecting ship position data that is consistent with the expected navigation mode of the target ship; connecting the optimal ship position data in each space-time area in chronological order to generate a navigation trajectory of the target ship.
[0060] The time axis during the navigation of the target ship is divided into several continuous time periods (for example, divided at fixed time intervals), and each time period corresponds to a space-time region; for each space-time region, its spatial range is defined as the maximum geographical range that the target ship may move within the time period (such as calculated by the maximum speed and time interval).
[0061] Furthermore, in each spatiotemporal region, a representative ship position data is selected from the fused ship position data as the optimal ship position data in the region. The selection criteria can refer to the following principles: Data quality: ship position data with higher indicators such as accuracy level, update frequency, and historical reliability in the data fusion process are preferred; Spatiotemporal consistency: ship position data with smooth transition between adjacent spatiotemporal regions and continuity greater than a predetermined continuity threshold are preferred to ensure the continuity of the trajectory; Motion mode matching: considering the expected navigation mode of the target ship (such as route planning, speed limit, etc.), select the ship position data that matches it. Then, the optimal ship position data in each spatiotemporal region are connected in chronological order to form a continuous navigation trajectory line.
[0062] It should be noted that when selecting the optimal ship position data in the spatiotemporal area, factors such as data quality, spatiotemporal consistency and motion pattern matching need to be considered.
[0063] In addition, the navigation track of the target ship can be visualized, stored or transmitted to the relevant supervision system.
[0064] It should be noted that, in this embodiment, the entire navigation process is divided into a plurality of continuous space-time regions according to the navigation time and speed of the target ship. These regions can be divided at fixed time intervals (such as every minute, every 5 minutes) or by navigation distance. In each space-time region, the optimal ship position data in the fused ship position data set is selected, and the optimal ship position data is the ship position data with the smallest positioning error, the highest confidence or in accordance with the prediction of a specific motion model. The selected optimal ship position data are connected in chronological order to form a complete navigation track of the target ship, and the longitude and latitude, timestamp and possible additional information (such as speed, heading) of each track point are saved.
[0065] Step S104, inputting the navigation track of the target ship into a pre-built normal ship trajectory model, calculating the trajectory feature similarity between the navigation track of the target ship and the navigation track of the normal ship, so as to identify the normal navigation behavior or abnormal navigation behavior of the target ship.
[0066] Optionally, step S104 includes: converting the navigation trajectory of the target ship into a feature representation consistent with a normal ship trajectory model, and calculating the trajectory feature similarity between the navigation trajectory characteristics of the target ship and the trajectory model of the normal ship using a distance measurement function or a machine learning model; comparing the trajectory feature similarity with a preset similarity threshold; if the trajectory feature similarity is greater than or equal to the similarity threshold, determining that the current navigation state of the target ship is in normal navigation behavior; if the trajectory feature similarity is less than the similarity threshold, determining that the current navigation state of the target ship has abnormal navigation behavior.
[0067] The above-mentioned normal ship trajectory data model is constructed by using machine learning or statistical analysis methods, such as cluster analysis, hidden Markov model, neural network, and is trained based on a large amount of historical trajectory data of normal sailing ships.
[0068] Optionally, a large amount of trajectory data of historical normal sailing ships is used to construct a model reflecting the characteristics of ship trajectory under normal sailing conditions through machine learning (such as cluster analysis, hidden Markov model, deep learning network, etc.) or statistical analysis methods; feature extraction is performed on the generated navigation trajectory of the target ship, including key features such as track shape, speed change, course change, docking behavior, turning characteristics, etc.; the trajectory characteristics of the target ship are compared with the normal ship trajectory data model, and the trajectory feature similarity is calculated using measurement methods such as cosine similarity, Euclidean distance, dynamic time warping (DTW), etc.; a threshold is set, and the navigation behavior of the target ship is judged according to the similarity score whether it belongs to the normal navigation range. If the similarity is lower than the threshold, it is considered that the target ship may have abnormal navigation behavior, such as deviation from the route, speeding, irregular track, etc.
[0069] Furthermore, the identified abnormal navigation behaviors are classified, such as deviation from the route, speeding, irregular track, too close safety distance, illegal turning, abnormal anchoring, etc., and the corresponding early warning mechanism is triggered to promptly notify the monitoring personnel or relevant regulatory departments; abnormal navigation events are recorded in detail, including the time and location of occurrence, information about the ships involved, type of abnormal navigation behavior, etc., and in-depth analysis is carried out in combination with background data such as weather, sea conditions, and traffic flow to identify potential risk factors or violation behavior patterns.
[0070] In view of the above problems, the present application proposes a method for identifying ship navigation behavior, and the specific technical effects are as follows:
[0071] (1) Multi-source ship position data fusion: By integrating ship position data from different signal sources such as AIS, radar, and Beidou, the positioning accuracy and data integrity are improved, providing a more accurate basis for subsequent navigation trajectory generation.
[0072] (2) Optimal ship position data screening and navigation trajectory generation: In each spatiotemporal region where the target ship is sailing, by setting reasonable screening rules (such as considering data freshness, source reliability, consistency with other data points, etc.), the optimal data points are selected from the fused ship position data to construct a continuous, smooth navigation trajectory that represents the actual navigation path. This method overcomes the local distortion or discontinuity that may be caused by a single data source, and the generated trajectory is more consistent with the actual movement law of the ship.
[0073] (3) Navigation behavior identification: Using a pre-built normal ship trajectory model (which may be based on statistical learning, machine learning or deep learning methods and trained on a large amount of normal navigation data), the similarity of trajectory features (such as distance, angle, speed change rate, turning radius, etc.) between the target ship's navigation trajectory and the typical normal trajectory in the model is calculated. By setting a threshold or using a classification algorithm, it can be determined whether the navigation behavior of the target ship is within the normal range. This improves the sensitivity and accuracy of abnormal navigation behavior detection, and helps to promptly detect potential dangerous operations, violations or security threats.
[0074] Optionally, the ship's automatic identification system (AIS) collects the target ship's position data: connecting to a global or regional AIS receiving network, or directly receiving real-time or historical ship position information from the AIS device installed on the target ship. These data usually include key information such as the ship's unique identifier (MMSI), geographic location (latitude and longitude), timestamp, speed, heading, etc.
[0075] Optionally, the radar system collects the position data of the target ship: the radar scans the target sea area and captures all echo signals including the target ship. The echo information is analyzed through signal processing technology to extract the position (latitude and longitude), relative speed, distance and other data of the target ship.
[0076] Optionally, satellite navigation system (e.g., Beidou Navigation System BDS) data access: If the target ship is equipped with a satellite navigation terminal, the positioning information sent by it can be received through a satellite communication link. This information usually contains precise geographic location (latitude and longitude), altitude, timestamp, etc.
[0077] Optionally, within the same time window, the ship position data from all signal sources are collected. These data should include location information (such as longitude and latitude) and other relevant parameters (such as timestamp, signal quality indicators, etc.); necessary pre-processing operations such as verification, denoising, and smoothing are performed on the collected data to ensure the quality and consistency of the data; the ship position data of all signal sources within the same time window are weighted averaged according to their respective weights. For the location information, the longitude and latitude can be weighted averaged independently; if other parameters are also included, they are also processed according to the corresponding weights; the result of the weighted average calculation is used as the fused ship position data.
[0078] In the embodiments of the present application, the accuracy, update frequency and historical reliability of different signal sources are taken into consideration, and the information from each source is balanced by weighted averaging to provide a comprehensive estimate of the target ship's position.
[0079] Optionally, the target ship's position data is represented as a state vector, which typically includes the ship's position (latitude and longitude) and speed (longitude speed, latitude speed), as well as other dynamic parameters that may be of interest (such as heading, speed); based on the first received position data, an initial value is assigned to the state vector, and a covariance matrix of the initial state is determined, which reflects an estimate of the uncertainty or error of the initial state.
[0080] Furthermore, a mathematical model describing the motion of the ship is constructed, such as the equation of motion based on Newtonian mechanics or a simplified dead reckoning model. The model should be able to predict the state at the next moment, given the current state and control input (such as rudder angle, engine power, etc.); an observation model is established for each signal source (AIS, radar, satellite navigation system), and the real state is mapped to the ship position data observed by the signal source. The observation model may include factors such as sensor characteristics (such as accuracy, delay), environmental influences (such as electromagnetic interference, multipath effect), etc.; corresponding noise covariance matrices are set for the system dynamic model and observation model respectively to reflect the deviation between the model prediction and the actual observation due to various uncertainties (such as wind and waves, equipment errors).
[0081] Furthermore, for each new time step (such as every second or every minute), the following iterative process is performed: the system dynamic model and the optimal estimated state at the previous moment are used to predict the state at the current moment, and the covariance of the predicted state is calculated based on the system noise covariance matrix; new ship position data from various signal sources are collected and necessary correction and standardization are performed; the Kalman gain is calculated based on the covariance of the predicted state, the observed noise covariance and the observation model, which is a coefficient that weighs the trust between the predicted value and the observed value; the optimal estimated state at the current moment is updated by weighted summation using the observed data, the Kalman gain and the predicted state; the covariance of the optimal estimated state at the current moment is updated based on the Kalman gain, the observed noise covariance and the predicted covariance; after each time step, the optimal estimated state at the current moment is output as the fused ship position data.
[0082] In the embodiment of the present application, information from different signal sources is subjected to a recursive update and prediction process of a Kalman filter, which effectively reduces noise, resolves data conflicts, and provides an optimal estimate of the position and speed of the target ship.
[0083] Optionally, based on the initial ship position data, a large number of particles P representing possible positions of the ship are generated. j (including longitude and latitude information), and assign an initial weight w j (0) (usually uniformly distributed); for each time step k, based on the ship motion model, the state of each particle is predicted and its position is updated; according to the new position of each particle, its likelihood corresponding to the observation data of each signal source is calculated (i.e., weight update); the weights of all particles are normalized to ensure that the total weight is 1; importance sampling is performed according to the particle weights, particles with larger weights are retained, and new particles are generated to replace particles with smaller weights, keeping the total number of particles unchanged; the particle data with the largest weight is taken as the fused ship position data output.
[0084] Optionally, after the particle filter is initialized (such as setting the number of particles, system model, etc.), for each new time step (such as every second or every minute), the following iterative process is performed: the particle state at the current moment is predicted using the system model and the particle state at the previous moment. At the same time, a prediction covariance is generated for each particle according to the system noise model (such as Gaussian distribution); new ship position data from various signal sources are collected and necessary correction and standardization are performed; the likelihood (or negative distance square) between each particle and the current observation data is calculated as the basis for updating its weight. The likelihood can be based on a distance function (such as Euclidean distance, Mahalanobis distance) or a more complex matching criterion; the weights of all particles are normalized to ensure that the total weight is 1, which is convenient for subsequent operations; according to the particle weight distribution, importance sampling is performed, particles with larger weights are retained, and new particles are generated to replace particles with smaller weights, keeping the total number of particles unchanged; the particle data with the largest weight is taken as the fused ship position data.
[0085] In the embodiment of the present application, information from different signal sources is recursively updated and predicted by a particle filter, which effectively reduces noise, resolves data conflicts, and provides an optimal estimate of the target ship's position.
[0086] Optionally, a large amount of historical navigation data of normal ships is collected, and the data is cleaned, sorted and standardized; cluster analysis, principal component analysis, deep learning and other methods are used to extract typical features and patterns of normal ship navigation trajectories, and a normal ship trajectory model is constructed; the navigation trajectory of the target ship is converted into a feature representation consistent with the normal ship trajectory model; a distance metric function (such as Euclidean distance, Mahalanobis distance, cosine similarity, etc.) or a machine learning model (such as support vector machine, neural network, etc.) is used to calculate the similarity between the navigation trajectory features of the target ship and the normal ship trajectory model.
[0087] For example, collect a large amount of historical trajectory data of normal sailing ships to ensure that the data covers various sailing conditions (such as weather, waterway, ship type, cargo type, etc.); clean the data, remove outliers and duplicate values, reasonably interpolate or fill in missing data, uniformly process timestamps, and convert all trajectories into a unified spatial coordinate system (such as WGS84); extract key features from each trajectory, such as speed, heading, sailing distance, voyage time, turning angle, docking time, sailing route, etc. More advanced dynamic features can also be extracted through time series analysis, spectrum analysis, curve fitting and other methods; select a suitable machine learning model (such as cluster analysis, hidden Markov model, deep learning network, etc.), and use the extracted features to train the model so that it can characterize the typical trajectory characteristics and patterns of normal sailing ships. During the model training process, feature selection, parameter tuning and other steps may be required; use an independent validation set or cross-validation to evaluate the model performance, adjust the model structure, parameters or feature selection according to the evaluation results, and ensure that the model has good generalization ability. Through the above implementation steps, a normal ship trajectory model can be constructed.
[0088] It should be noted that the above-mentioned normal ship trajectory model is based on a large amount of normal ship historical navigation data and extracts its typical features; wherein, one or more methods of cluster analysis, principal component analysis or deep learning are adopted when constructing the normal ship trajectory model.
[0089] For example, the navigation trajectory of the target ship is subjected to the same data preprocessing and feature extraction as when the model is constructed, ensuring that it is consistent with the model input format; the trajectory features of the target ship are converted into feature vectors that are compatible with the normal ship trajectory model, ensuring that the two can perform effective similarity calculations; an appropriate similarity measurement method (such as Euclidean distance, Mahalanobis distance, cosine similarity, dynamic time warping, etc.) is used to calculate the similarity between the feature vector of the target ship and the normal ship trajectory model; based on the experience accumulated during the model training and verification process, or industry standards and regulatory requirements, one or more similarity thresholds are set to distinguish between normal navigation and abnormal navigation; the calculated similarity is compared with the set threshold: if the similarity is higher than the threshold, it is determined that the target ship exhibits normal navigation behavior; if the similarity is lower than the threshold, it is determined that the target ship may have abnormal navigation behavior, and further analysis or corresponding measures are required.
[0090] In an embodiment of the present application, the above-mentioned implementation method can input the navigation trajectory of the target ship into a pre-constructed normal ship trajectory model, calculate the similarity of its trajectory features with those of a normal ship, and identify its normal navigation behavior or abnormal navigation behavior accordingly.
[0091] As an optional embodiment, the above method also includes: obtaining performance indicators of different signal sources, wherein the performance indicators include at least one of the following: signal strength, data update frequency, and noise level; inputting the performance indicators of different signal sources into a pre-constructed weight prediction model to obtain real-time weights of different signal sources; wherein the weight prediction model is obtained by training with a neural network or a machine learning algorithm using multiple first training data, and the first training data is derived from the performance data of historical signal sources and the weights set by expert experience; according to the real-time weights of different signal sources, data fusion is performed on the ship position data within the same time window to obtain fused ship position data.
[0092] Optionally, the signal strength, data update frequency, noise level and other performance indicators of each signal source (AIS, radar, satellite navigation system) are monitored in real time; a weight prediction model is trained using a neural network or other machine learning algorithm, with the input being the performance indicators of each signal source and the output being the corresponding weight value. The model training data is derived from the historical signal source performance data and the weights set by expert experience; in data fusion, the performance indicators of each signal source currently monitored are input into the weight prediction model in real time to obtain the real-time weight of each signal source; based on the real-time weight of each signal source, the ship position data in the same time window is weighted averaged or a similar fusion algorithm is used to generate the fused ship position data.
[0093] In the embodiment of the present application, the data fusion process also includes adaptive weight allocation, that is, using a machine learning algorithm to dynamically adjust the weight of each signal source's contribution to the final ship position data according to its real-time performance. In other words, the weight of each signal source's contribution to the final ship position data can be dynamically adjusted according to its real-time performance, for example, considering factors such as signal strength, data update frequency, noise level, etc., using a machine learning algorithm (such as a neural network) training model to predict and allocate weights.
[0094] As an optional embodiment, the above method also includes: obtaining multi-source information of the target ship; wherein the multi-source information includes the target ship's position data, ship communication data, navigation log and meteorological data; inputting the multi-source information of the target ship into a trained ship navigation behavior recognition model to obtain the current navigation behavior category and probability of the target ship; wherein the ship navigation behavior recognition model is obtained by training a deep learning model using multiple second training data, and the second training data includes multi-source information and corresponding known navigation behavior labels.
[0095] Optionally, multi-source information such as the target ship's position data, ship communication data, navigation log, meteorological data, etc. is collected to ensure the time synchronization and unified format of the information; a suitable deep learning model (such as CNN, RNN, GNN) is selected according to the application scenario, and the input layer, hidden layer, and output layer structure are designed. The model is trained to identify the ship's navigation behavior, and the training data includes multi-source information and corresponding known navigation behavior labels; in the behavior recognition step, the multi-source information of the target ship is input into the trained ship navigation behavior recognition model, and the model output is the current navigation behavior category and probability of the ship.
[0096] In an embodiment of the present application, in the process of behavior recognition, multimodal fusion technology can also be used to combine multi-source information such as ship position data, ship communication data, navigation logs, meteorological data, etc., and use deep learning models to perform behavior recognition. Not only does it use ship position data, but it also integrates multiple information sources such as ship communication data, navigation logs, meteorological data, etc., and uses deep learning models (such as convolutional neural networks, recurrent neural networks, or graph neural networks) to perform multimodal fusion analysis to more accurately identify complex navigation behaviors.
[0097] Optionally, the device signals and base station coverage area of the target ship are monitored in real time, and when the target ship enters from one coverage area into another overlapping coverage area, the change is automatically detected; in the overlapping coverage area, the most suitable signal source for continued tracking is automatically selected based on factors such as the signal strength, data quality, and availability of each signal source; the received new signal source data is smoothed to ensure seamless connection with the previous signal source data to avoid trajectory interruption or mutation due to signal source switching; the continuity and rationality of the trajectory during signal source switching is ensured by checking the time interval, position change and other information of several consecutive trajectory points, and interpolation correction is performed when necessary.
[0098] It should be noted that the methods for detecting changes in signal coverage areas include but are not limited to signal strength monitoring, GPS coordinate change detection or network registration information change monitoring. The basis for selecting the best signal source includes but is not limited to signal strength, data quality score, historical reliability statistics, and current network load. The means of data smooth transition include but are not limited to linear interpolation, polynomial fitting, Kalman filtering or particle filtering.
[0099] In an embodiment of the present application, in the overlapping coverage area of the equipment signal of the target ship and the base station, tracking and handover are automatically performed without manual intervention, thereby ensuring seamless continuity of the ship's trajectory.
[0100] Optionally, the navigation status (speed, heading, navigation area) of the target ship, as well as future route predictions, current weather, sea conditions and other environmental information are obtained in real time; a machine learning algorithm (such as decision tree, reinforcement learning) is used to train a signal source switching decision model, with the input being the navigation status, future route predictions, environmental information, etc., and the output being whether signal source switching is required and which signal source to switch to; in the automatic tracking and handover of the signal source, the current navigation status, future route predictions, and environmental information are input into the signal source switching decision model in real time, and based on the model output results, if it is predicted that signal quality problems may occur in the future, the signal source is switched in advance.
[0101] In the embodiment of the present application, in the process of automatic tracking and handover of the signal source, an intelligent switching strategy is also included. The strategy predicts situations that may affect the signal quality according to the navigation status of the ship, the prediction of future routes, and environmental factors, and switches the signal source in advance. In other words, in addition to selecting the best signal source, an intelligent switching strategy is introduced to predict situations that may affect the signal quality according to the navigation status of the ship (such as speed, heading, navigation area), future route prediction, and environmental factors such as weather and sea conditions, and switch the signal source in advance to optimize the trajectory continuity.
[0102] The following is a detailed description in conjunction with another embodiment.
[0103] Embodiment 2
[0104] A ship navigation behavior identification system provided in this embodiment includes multiple implementation units, each implementation unit corresponds to each implementation step in the above-mentioned embodiment 1.
[0105] Figure 2 is a schematic diagram of an optional ship navigation behavior recognition system according to an embodiment of the present invention, such as Figure 2 As shown, the ship navigation behavior identification system may include: a ship position data acquisition unit 21, a data fusion unit 22, a navigation track generation unit 23, and a navigation behavior determination module 24.
[0106] The ship position data acquisition unit 21 is used to acquire the ship position data of the target ship collected by N signal sources, the signal sources include: a ship automatic identification system, a radar system and a satellite navigation system, and N is a positive integer greater than 1.
[0107] The data fusion unit 22 is used to fuse the position data of the target ship under different signal sources to obtain fused position data.
[0108] The navigation track generating unit 23 is used to select the optimal ship position data from the fused ship position data in each time-space area where the target ship is sailing, and generate the navigation track of the target ship based on the optimal ship position data.
[0109] The navigation behavior identification module 24 is used to input the navigation track of the target ship into a pre-built normal ship track model, calculate the track feature similarity between the navigation track of the target ship and the navigation track of the normal ship, so as to identify the normal navigation behavior or abnormal navigation behavior of the target ship.
[0110] The above-mentioned ship navigation behavior identification system can be used to obtain the position data of the target ship collected by N signal sources through a ship position data acquisition unit, the signal sources include: a ship automatic identification system, a radar system and a satellite navigation system, and N is a positive integer greater than 1; a data fusion unit, used to fuse the position data of the target ship under different signal sources to obtain fused position data; a navigation trajectory generation unit, used to screen out the optimal ship position data from the fused ship position data in each time and space area where the target ship navigates, and generate the navigation trajectory of the target ship based on the optimal ship position data; a navigation behavior discrimination module, used to input the navigation trajectory of the target ship into a pre-constructed normal ship trajectory model, calculate the trajectory feature similarity between the navigation trajectory of the target ship and the navigation trajectory of the normal ship, so as to identify the normal navigation behavior or abnormal navigation behavior of the target ship. In this embodiment, a pre-built normal ship trajectory model is used to calculate the trajectory feature similarity between the target ship's navigation trajectory and the typical normal trajectory in the model; by setting a threshold or using a classification algorithm, it can be judged whether the navigation behavior of the target ship is within the normal range, thereby improving the sensitivity and accuracy of abnormal navigation behavior detection, helping to timely discover potential dangerous operations, violations or security threats, and thus solving the technical problem of the inability to accurately identify the navigation behavior of ships in related technologies.
[0111] Optionally, the data fusion unit includes: a preprocessing module, used to perform data preprocessing on the position data of the target ship under different signal sources to obtain preprocessed position data, wherein the data preprocessing includes at least one of the following: removing invalid data, removing duplicate data, removing erroneous data, and converting the position data of different signal sources into a unified data format, and time aligning the position data of different signal sources; a data fusion module, used to use a data fusion algorithm to perform fusion calculation on the preprocessed position data to obtain fused position data.
[0112] Optionally, the data fusion module includes: a weight parameter assignment submodule, which is used to assign corresponding weight parameters to the ship position data of different signal sources according to the accuracy level, update frequency and historical reliability of different signal sources, and for the ship position data of different signal sources in the same time window, weighted average calculation is performed according to their weights, and the calculated weighted average data is used as the fused ship position data; and / or, a ship position data update submodule, which is used to recursively update the ship position data of different signal sources using a Kalman filter, and use the output optimal estimation data as the fused ship position data; and / or, a ship position fusion submodule, which is used to adopt a particle filtering algorithm, simulate a particle swarm to characterize the position state of the ship, update the particle weight distribution according to the ship position data of different signal sources, and use the particle data with the largest weight as the fused ship position data.
[0113] Optionally, the navigation trajectory generating unit includes: a time axis division module, used to divide the time axis during the navigation of the target ship into multiple continuous time periods, each time period corresponds to a space-time area, wherein the space-time area represents the maximum geographical range that the target ship can move within the corresponding time period; a position data selection module, used to select one of the position data from the fused position data in each space-time area based on a predetermined reference standard as the optimal position data for the space-time area, wherein the predetermined reference standard includes at least one of the following: selecting position data whose accuracy level, update frequency, and historical reliability index in the data fusion process are greater than the index threshold, selecting position data with smooth transition between adjacent space-time areas and continuity greater than a predetermined continuity threshold, and selecting position data that is consistent with the expected navigation mode of the target ship; a navigation trajectory generating module, used to connect the optimal position data in each space-time area in chronological order to generate the navigation trajectory of the target ship.
[0114] Optionally, the navigation behavior discrimination module includes: a trajectory feature similarity calculation submodule, which is used to convert the navigation trajectory of the target ship into a feature representation consistent with the normal ship trajectory model, and use a distance measurement function or a machine learning model to calculate the trajectory feature similarity between the navigation trajectory characteristics of the target ship and the trajectory model of the normal ship; a similarity comparison submodule, which is used to compare the trajectory feature similarity with a preset similarity threshold; a first navigation behavior discrimination submodule, which is used to determine that the current navigation state of the target ship is in normal navigation behavior when the trajectory feature similarity is greater than or equal to the similarity threshold; and a second navigation behavior discrimination submodule, which is used to determine that the current navigation state of the target ship has abnormal navigation behavior when the trajectory feature similarity is less than the similarity threshold.
[0115] Optionally, the ship navigation behavior identification system also includes: a performance indicator acquisition unit, used to obtain performance indicators of different signal sources, wherein the performance indicators include at least one of the following: signal strength, data update frequency, noise level; a signal source weight determination unit, used to input the performance indicators of different signal sources into a pre-constructed weight prediction model to obtain real-time weights of different signal sources, wherein the weight prediction model is obtained by training with a neural network or a machine learning algorithm using multiple first training data, and the first training data is derived from the performance data of historical signal sources and the weights set by expert experience; a ship position data fusion unit, used to perform data fusion on the ship position data within the same time window according to the real-time weights of different signal sources to obtain fused ship position data.
[0116] Optionally, the ship navigation behavior identification system also includes: a multi-source information acquisition unit, used to acquire multi-source information of the target ship, wherein the multi-source information includes the target ship's position data, ship communication data, navigation log and meteorological data; a navigation behavior identification unit, used to input the multi-source information of the target ship into a trained ship navigation behavior identification model to obtain the current navigation behavior category and probability of the target ship, wherein the ship navigation behavior identification model is trained by a deep learning model using multiple second training data, and the second training data includes multi-source information and corresponding known navigation behavior labels.
[0117] The above-mentioned ship navigation behavior identification system may also include a processor and a memory. The above-mentioned ship position data acquisition unit 21, data fusion unit 22, navigation trajectory generation unit 23, navigation behavior judgment module 24, etc. are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize the corresponding functions.
[0118] The processor includes a kernel, which retrieves the corresponding program unit from the memory. One or more kernels can be set, and the ship's navigation behavior can be monitored to see if it is abnormal by adjusting kernel parameters.
[0119] The above-mentioned memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash memory (fl ash RAM), and the memory includes at least one storage chip.
[0120] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the method for identifying the ship navigation behavior of any one of the above-mentioned embodiments one.
[0121] According to another aspect of an embodiment of the present invention, there is also provided an electronic device, comprising one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by one or more processors, the one or more processors implement the method for identifying the ship navigation behavior of any one of the above-mentioned embodiments one.
[0122] The present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the method for identifying the navigation behavior of a ship described in each embodiment of the present application.
[0123] The present application also provides a computer program product, including a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method for identifying the navigation behavior of a ship described in each embodiment of the present application are implemented.
[0124] Figure 3 1 is a hardware structure block diagram of an electronic device (or mobile device) for a method for identifying ship navigation behavior according to an embodiment of the present invention. Figure 3 As shown, the electronic device may include one or more ( Figure 3 302a, 302b, ..., 302n are used to illustrate) processor (the processor may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 304 for storing data. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the I / O interface), a network interface, a keyboard, a power supply and / or a camera. It can be understood by those skilled in the art that Figure 3 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 3 More or fewer components as shown, or with Figure 3 Different configurations are shown.
[0125] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.
[0126] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.
[0127] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.
[0128] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.
[0129] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.
[0130] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of 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, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.
[0131] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.
Claims
1. A method for identifying ship navigation behavior, characterized in that: include: Obtaining the position data of the target ship collected by N signal sources, the signal sources including: a ship automatic identification system, a radar system and a satellite navigation system, where N is a positive integer greater than 1; fusing the position data of the target ship under different signal sources to obtain fused position data; In each spatiotemporal region where the target ship is sailing, selecting optimal ship position data from the fused ship position data, and generating a navigation track of the target ship based on the optimal ship position data; The navigation track of the target ship is input into a pre-built normal ship track model, and the track feature similarity between the navigation track of the target ship and the navigation track of the normal ship is calculated to identify the normal navigation behavior or abnormal navigation behavior of the target ship.
2. The identification method according to claim 1, characterized in that: The ship position data of the target ship under different signal sources are fused to obtain fused ship position data, including: Performing data preprocessing on the position data of the target ship under different signal sources to obtain preprocessed position data, wherein the data preprocessing includes at least one of the following: removing invalid data, removing duplicate data, removing erroneous data, and converting the position data of different signal sources into a unified data format, and performing time alignment on the position data of different signal sources; A data fusion algorithm is used to perform fusion calculation on the preprocessed ship position data to obtain the fused ship position data.
3. The identification method according to claim 2, characterized in that: Using a data fusion algorithm to perform fusion calculation on the pre-processed ship position data to obtain the fused ship position data includes: Assigning corresponding weight parameters to the ship position data of different signal sources according to the accuracy level, update frequency and historical reliability of different signal sources, performing weighted average calculation according to the weights of the ship position data of different signal sources in the same time window, and using the calculated weighted average data as the fused ship position data; and / or, recursively updating the ship position data from different signal sources using a Kalman filter, and using the outputted optimal estimation data as the fused ship position data; and / or, The particle filter algorithm is adopted to characterize the position state of the ship by simulating a particle swarm, and the particle weight distribution is updated according to the ship position data of different signal sources, and the particle data with the largest weight is used as the fused ship position data.
4. The identification method according to claim 1, characterized in that: In each spatiotemporal region where the target ship is sailing, the step of selecting optimal ship position data from the fused ship position data and generating the navigation track of the target ship based on the optimal ship position data comprises: Divide the time axis during the navigation of the target ship into a plurality of continuous time periods, each time period corresponds to a time-space region, wherein the time-space region represents the maximum geographical range within which the target ship can move within the corresponding time period; In each of the spatiotemporal regions, one of the ship position data is selected from the fused ship position data based on a predetermined reference standard as the optimal ship position data for the spatiotemporal region, wherein the predetermined reference standard includes at least one of the following: selecting ship position data whose accuracy level, update frequency, and historical reliability index are greater than an index threshold during data fusion, selecting ship position data with a smooth transition between adjacent spatiotemporal regions and a continuity greater than a predetermined continuity threshold, and selecting ship position data that is consistent with the expected navigation mode of the target ship; The optimal ship position data in each of the time-space regions are connected in time sequence to generate the navigation track of the target ship.
5. The identification method according to claim 1, characterized in that: Inputting the navigation track of the target ship into a pre-built normal ship track model, calculating the track feature similarity between the navigation track of the target ship and the navigation track of the normal ship, so as to identify the normal navigation behavior or abnormal navigation behavior of the target ship, including: The navigation track of the target ship is converted into a feature representation consistent with the normal ship track model, and the track feature similarity between the navigation track feature of the target ship and the track model of the normal ship is calculated using a distance metric function or a machine learning model; The trajectory feature similarity is compared with a preset similarity threshold; if the trajectory feature similarity is greater than or equal to the similarity threshold, it is determined that the current navigation state of the target ship is in normal navigation behavior; if the trajectory feature similarity is less than the similarity threshold, it is determined that the current navigation state of the target ship has abnormal navigation behavior.
6. The identification method according to claim 1, characterized in that: Also includes: Acquire performance indicators of different signal sources, wherein the performance indicators include at least one of the following: signal strength, data update frequency, and noise level; Inputting the performance indicators of different signal sources into a pre-built weight prediction model to obtain real-time weights of different signal sources, wherein the weight prediction model is obtained by training a neural network or a machine learning algorithm using a plurality of first training data, and the first training data is derived from the performance data of historical signal sources and the weights set by expert experience; According to the real-time weights of different signal sources, the ship position data in the same time window are fused to obtain the fused ship position data.
7. The identification method according to claim 1, characterized in that: Also includes: Acquire multi-source information of the target ship, wherein the multi-source information includes position data, ship communication data, navigation log and meteorological data of the target ship; The multi-source information of the target ship is input into a trained ship navigation behavior recognition model to obtain the current navigation behavior category and probability of the target ship, wherein the ship navigation behavior recognition model is trained by a deep learning model using multiple second training data, and the second training data includes multi-source information and corresponding known navigation behavior labels.
8. A system for identifying ship navigation behavior, characterized in that: include: A ship position data acquisition unit, used to acquire the ship position data of the target ship collected by N signal sources, the signal sources include: a ship automatic identification system, a radar system and a satellite navigation system, and N is a positive integer greater than 1; A data fusion unit, used for fusing the position data of the target ship under different signal sources to obtain fused position data; A navigation track generating unit, configured to select optimal ship position data from the fused ship position data in each spatiotemporal region where the target ship is sailing, and generate a navigation track of the target ship based on the optimal ship position data; The navigation behavior identification module is used to input the navigation track of the target ship into a pre-built normal ship track model, calculate the track feature similarity between the navigation track of the target ship and the navigation track of the normal ship, so as to identify the normal navigation behavior or abnormal navigation behavior of the target ship.
9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the method for identifying the navigation behavior of a ship as described in any one of claims 1 to 7.
10. An electronic device, characterized in that: It includes one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method for identifying the navigation behavior of a ship as described in any one of claims 1 to 7.
Citation Information
Patent Citations
Ship anomaly detection method based on improved trajectory segment DBSCAN clustering
CN113032502A
Ship track behavior anomaly detection method, system and device and storage medium
CN116608861A
Ship early warning method and system based on data fusion
CN117992906A
Ship image trajectory tracking and prediction method based on ship heading recognition
US20240013402A1
Cited By
Rim motor cooperative control method and system
CN119182313A
A rim motor coordinated control method and system
CN119182313B
Method and device for determining marine photovoltaic ship navigation data, medium and equipment
CN119580050A
Method and Device for Determining Ship Navigation Data of Offshore Photovoltaic, Medium and Equipment
CN119580050B
Ship radar identification method based on deep learning
CN119780871A
