Ship abnormal behavior analysis method and system

Abstract

The invention relates to the technical field of ship behavior analysis, and discloses a ship abnormal behavior analysis method and system. According to the ship behavior prediction analysis model, behavior information of the target ship at the next moment is predicted; comparing the prediction result with a ship behavior threshold value to obtain an initial analysis result of the abnormal ship behavior; the behavior stability of the target ship and the prediction accuracy of the ship behavior prediction analysis model are obtained; and by combining the behavior stability of the target ship and the accuracy of the ship behavior prediction analysis model, adjusting the initial analysis result, and obtaining a final analysis result of the ship abnormal behavior. The problem that abnormal behavior analysis only depends on a prediction result is solved, and the accuracy of an abnormal behavior judgment result is improved by considering multiple aspects.

Description

Technical Field

[0001] The present invention relates to the technical field of ship behavior analysis, and in particular to a method and system for analyzing abnormal ship behavior. Background Art

[0002] Abnormal ship behavior usually refers to behavior that deviates from the normal behavior pattern of the ship. This deviation may be caused by a variety of factors, such as environmental uncertainty, uncertainty in crew operations, and problems in information acquisition or identification. Analysis of abnormal ship behavior can not only capture and identify non-standard or potentially dangerous behaviors of ships during navigation on the water in real time, thereby effectively preventing water traffic accidents and ensuring the safety of personnel, cargo and water environment; it can also optimize waterway use, improve shipping efficiency, and reduce unnecessary delays and costs by analyzing the reasons behind abnormal behavior. In addition, this analysis process helps to ensure that all ships comply with relevant laws and regulations and maintain the legal order of water traffic. Therefore, analysis of abnormal ship behavior is an important means to improve the level of water traffic management, ensure safety and efficiency, and is of great significance to promoting the sustainable development of the shipping industry.

[0003] At present, when analyzing ship anomalies, the prediction results of ship behavior are usually compared with pre-set thresholds. When the predicted ship behavior parameters exceed these thresholds, the system will trigger an alarm, indicating that abnormal behavior may exist. However, the accuracy of the prediction results directly determines the effectiveness of abnormal behavior analysis. If there are deviations or errors in the prediction results themselves, then the threshold comparison based on these prediction results will also lose its meaning, greatly reducing the accuracy of the ship abnormal behavior analysis results.

[0004] Therefore, a more accurate and comprehensive method for analyzing abnormal ship behavior has become an urgent need to improve the accuracy and efficiency of identifying abnormal ship behavior and provide more powerful protection for water traffic safety. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides a method for analyzing abnormal behavior of a ship, the method comprising the following steps: S1: Obtain the current ship data and ocean hydrological and meteorological environmental data of the target ship, input the trained ship behavior prediction analysis model, and predict the behavior information of the target ship at the next moment; S2: Compare the prediction results in S1 with the ship behavior threshold to obtain the initial analysis results of the abnormal ship behavior; S3: The behavior stability of the target ship is obtained by calculating the historical behavior information of the target ship; S4: Calculate the prediction accuracy of the ship behavior prediction analysis model at the current moment; S5: Combine the behavior stability of the target ship and the prediction accuracy of the ship behavior prediction analysis model to judge the initial analysis results, so as to determine the final analysis results of the abnormal behavior of the ship.

[0006] Furthermore, the ship behavior prediction and analysis model includes a GRU unit, a Spatial-ReductionAttention unit and a fully connected layer.

[0007] Furthermore, the behavior information includes heading, speed, and longitude and latitude.

[0008] Furthermore, in S3, the historical behavior information of the target ship is calculated to obtain the behavior stability of the target ship, which is specifically: S31: collecting historical behavior information of the target ship within a certain time range, and preprocessing the collected historical behavior information; S32: Calculate the heading, speed, and longitude and latitude in the pre-processed historical behavior information to obtain heading stability, speed stability, and longitude and latitude stability; S33: Based on the weighted method, the behavior stability of the target ship is obtained according to the heading stability, speed stability and longitude and latitude stability.

[0009] Furthermore, in S5, the initial analysis result is judged in combination with the behavior stability of the target ship and the prediction accuracy of the ship behavior prediction analysis model, so as to determine the final analysis result of the abnormal behavior of the ship, specifically: Constructing a feature vector according to the behavior stability of the target ship and the prediction accuracy of the ship behavior prediction analysis model; According to the characteristic vector, the model is adjusted by analyzing the result to obtain the confidence of the initial analysis result; According to the confidence level and in combination with the initial analysis result, a final analysis result of the abnormal behavior of the ship is determined.

[0010] Further, according to the characteristic vector, the model is adjusted by analyzing the result to obtain the confidence of the initial analysis result, including: Normalizing the feature vector to obtain a normalized feature vector; According to the normalized feature vector, the model is adjusted by the trained analysis result to obtain a decision value corresponding to the initial analysis result; The decision value is converted to serve as the confidence level of the initial analysis result.

[0011] Furthermore, an activation function is used to convert the decision value.

[0012] On the other hand, the present application also provides a ship abnormal behavior analysis system, which executes any of the above-mentioned ship abnormal behavior analysis methods, and the system includes: an initial analysis result acquisition module, a ship behavior stability acquisition module, a model accuracy acquisition module, and an output module; The initial analysis result acquisition module is used to obtain the ship data and ocean hydrological and meteorological environmental data of the target ship at the current moment, input the trained ship behavior prediction analysis model, and predict the behavior information of the target ship at the next moment; compare the prediction result with the ship behavior threshold to obtain the initial analysis result of the abnormal behavior of the ship; The ship behavior stability acquisition module obtains the behavior stability of the target ship by calculating the historical behavior information of the target ship; The model accuracy acquisition modulus is used to count the prediction accuracy of the ship behavior prediction analysis model at the current moment; The output module is used to judge the initial analysis results by combining the behavior stability of the target ship and the prediction accuracy of the ship behavior prediction analysis model, so as to determine the final analysis results of the abnormal behavior of the ship.

[0013] The embodiments of the present invention have the following technical effects: In the ship abnormal behavior analysis method provided by the present application, the behavior information of the target ship at the next moment is predicted based on the ship behavior prediction analysis model; the prediction result is compared with the ship behavior threshold to obtain the initial analysis result of the ship abnormal behavior; and then the initial analysis result is judged by combining the behavior stability of the target ship and the accuracy of the ship behavior prediction analysis model to determine the final analysis result of the ship abnormal behavior. This overcomes the problem of completely relying on the prediction results for analysis when analyzing abnormal behavior, and avoids the influence of the deviation of the prediction results themselves on the analysis results.

[0014] Among them, when judging the initial analysis results, not only the prediction accuracy of the ship behavior prediction analysis model that obtains the prediction results is considered, but also the behavioral stability of the target ship is taken into account. The prediction results and the behavioral characteristics of the target ship itself are combined to further improve the accuracy of the ship abnormal behavior analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1This is a flow chart of a method for analyzing abnormal behavior of a ship provided by an embodiment of the present invention; Figure 2 The present invention provides a simplified structural diagram of a ship abnormal behavior analysis system. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be described clearly and completely below. 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 belong to the scope of protection of the present invention.

[0018] A method for analyzing abnormal behavior of a ship, such as Figure 1 As shown, the method comprises the following steps: S1: Obtain the current ship data and ocean hydrological and meteorological environmental data of the target ship, input the trained ship behavior prediction analysis model, and predict the behavior information of the target ship at the next moment; In some embodiments, the behavior information includes heading, speed, and longitude and latitude.

[0019] In some embodiments, the ship behavior prediction analysis model includes a GRU unit, a Spatial-ReductionAttention unit and a fully connected layer.

[0020] Aiming at the single ship prediction in complex environment, a bidirectional gated recurrent neural network (SRA-BiGRU) integrating spatial reduction attention mechanism is proposed. Compared with the single model structure, this model utilizes the temporal feature ability of GRU unit learning data and captures the key factors affecting ship behavior by introducing Spatial-Reduction Attention unit, which reduces the influence of interference information while retaining important features, so that the network can process longer temporal information, making the model robust and accurate, and has the characteristics of high computational efficiency and strong engineering applicability.

[0021] Specifically, ship data and ocean hydrological and meteorological environmental data are used as input parameters of the ship behavior prediction and analysis model. The GRU unit is used to obtain feature vectors with time series characteristics. The Spatial-Reduction Attention mechanism unit obtains the weight coefficients corresponding to each feature vector with time series characteristics through the attention mechanism. This enables rapid identification of key variables related to ship trajectory prediction, such as ocean information, and accurate extraction of key information affecting the current ship's navigation. At the same time, it can capture time dependencies and identify hidden key factors from historical ship trajectory data. It can extract key time information that affects the current ship's trajectory, and dynamically extract the corresponding key influencing factors by adjusting the weight coefficients, accurately predicting the target ship's behavior information at the next moment.

[0022] S2: Compare the prediction results in S1 with the ship behavior threshold to obtain the initial analysis results of the abnormal ship behavior; According to the historical behavior data of the ship, the heading and speed of the ship at each position (expressed in longitude and latitude) under normal circumstances are set as the ship behavior threshold. The heading, speed, longitude and latitude of the target ship at the next moment are predicted through the ship behavior prediction and analysis model. The predicted heading, speed, longitude and latitude are compared with the ship behavior threshold respectively: if the predicted longitude and latitude are not within the threshold, the ship behavior is judged to be abnormal; if the predicted heading deviates from the threshold, the ship behavior is judged to be abnormal; if the predicted speed is not within the threshold, the speed is too high or too low, the ship behavior is judged to be abnormal.

[0023] S3: The behavior stability of the target ship is obtained by calculating the historical behavior information of the target ship; S31: collecting historical behavior information of the target ship within a certain time range, and preprocessing the collected historical behavior information; Collect historical behavior information, that is, collect the historical behavior information of the target ship in the past period of time, including data such as course, speed, longitude and latitude. These data can be obtained from the ship's automatic identification system (AIS), navigation log, monitoring system and other channels to ensure the integrity, accuracy and consistency of the data.

[0024] The collected data also needs to be preprocessed, including data cleaning: removing duplicate, erroneous or abnormal data points, such as sudden changes in heading, abnormal speed, etc.; interpolation: for missing data points, use appropriate interpolation methods (such as linear interpolation, neighboring point interpolation, etc.) to fill them.

[0025] S32: Calculate the heading, speed, and longitude and latitude in the pre-processed historical behavior information to obtain heading stability, speed stability, and longitude and latitude stability; Heading stability: The standard deviation is used to measure the stability of the heading. The smaller the standard deviation, the more stable the heading.

[0026] ; Among them, A represents heading stability, represents the data point of the i-th heading, represents the average value of the heading data, and n represents the number of data points.

[0027] Speed ​​stability: Standard deviation is also used to measure speed stability B.

[0028] Longitude and latitude stability C: The stability of longitude and latitude is measured by the position change rate. The smaller the change rate, the more stable the position, that is, the vehicle is driving smoothly along the established route. The position change rate can be expressed by the ratio of the direct distance between two consecutive data points to the time interval between the two data points, where the direct distance between two consecutive data points can be calculated using longitude and latitude.

[0029] S33: Based on the weighted method, the behavior stability of the target ship is obtained according to the heading stability, speed stability and latitude and longitude stability. The weight coefficients corresponding to the heading stability, speed stability and latitude and longitude stability can be set according to the fluctuation size of each in the historical data of the target ship, or can be predicted by a neural network model, or obtained by other methods, which are not limited here.

[0030] S4: Calculate the prediction accuracy of the ship behavior prediction analysis model at the current moment; The prediction accuracy can be expressed by the ratio of the number of accurate predictions by the ship behavior prediction analysis model to the total number of predictions. The behavior information of the ship predicted by the ship behavior prediction analysis model is compared with the behavior information of the ship acquired in real time. If the deviation is within a predetermined range, the prediction is determined to be accurate. Then, the number of accurate predictions and the total number of predictions by the ship behavior prediction analysis model are counted.

[0031] S5: Combine the behavior stability of the target ship and the prediction accuracy of the ship behavior prediction analysis model to determine the initial analysis results, so as to determine the final analysis results of the abnormal behavior of the ship, specifically: S51: constructing a feature vector according to the behavior stability of the target ship and the prediction accuracy of the ship behavior prediction analysis model; Exemplarily, a feature vector in the form of a two-dimensional array is constructed, where each feature vector contains two elements: a behavior stability value (y1) and a prediction accuracy value (y2).

[0032] Prepare label data, that is, the true label of whether the ship behavior corresponding to each feature vector is abnormal (1 indicates abnormality and 0 indicates normality).

[0033] S52: According to the feature vector, the model is adjusted by analyzing the results to obtain the confidence of the initial analysis result, which specifically includes: S521: normalizing the feature vector to obtain a normalized feature vector to ensure that different features are comparable in value.

[0034] S522: According to the normalized feature vector, the model is adjusted by the trained analysis result to obtain a decision value corresponding to the initial analysis result; For example, support vector machine (SVM) can be selected as the algorithm for adjusting the model of analysis results. Based on its powerful classification and regression functions, in the analysis of abnormal ship behavior, SVM can learn a decision boundary based on the two characteristics of behavior stability and prediction accuracy, thereby distinguishing normal behavior from abnormal behavior and giving the confidence of the analysis results.

[0035] SVM does not directly output confidence, but confidence can be indirectly evaluated by observing the decision function (that is, how close the model output is to the decision boundary). Specifically, the output value of the decision function (also called the decision value or score) can be used as a measure of confidence. The larger the decision value (for abnormal behavior) or the smaller it is (for normal behavior), the higher the confidence of the model in the prediction result.

[0036] The prepared data set is divided into a training set and a test set for training and validating the trained SVM model. The trained analysis results are obtained to adjust the model.

[0037] S523: converting the decision value to serve as the confidence level of the initial analysis result.

[0038] In some embodiments, an activation function is used to transform the decision value.

[0039] In order to convert the decision value into confidence, a sigmoid activation function or other mapping functions can be used for conversion, but it should be noted that this conversion is usually based on experience or specific application scenarios and there is no fixed formula. In practical applications, the conversion rules can be customized as needed.

[0040] The confidence can be presented in the form of probability, that is, the probability that the ship's behavior is abnormal (or normal). However, it should be noted that this interpretation is based on the confidence output by the SVM model and specific conversion rules, and is not an absolute probability value.

[0041] Combination of classification labels and confidence levels: When outputting classification labels (abnormal or normal), the corresponding confidence levels are also included to provide a more comprehensive understanding of the reliability of the analysis results and to adjust the model analysis results.

[0042] S53: Determine the final analysis result of the abnormal behavior of the ship based on the confidence level and the initial analysis result.

[0043] The reliability of the initial analysis results can be determined based on the converted confidence values. For example, a confidence threshold can be set, and when the confidence exceeds the threshold, the analysis results are considered reliable, otherwise it is considered that there may be errors or further verification is required.

[0044] In the ship abnormal behavior analysis method provided by the present application, the behavior information of the target ship at the next moment is predicted based on the ship behavior prediction analysis model; the prediction result is compared with the ship behavior threshold to obtain the initial analysis result of the ship abnormal behavior; and then the initial analysis result is judged by combining the behavior stability of the target ship and the accuracy of the ship behavior prediction analysis model to determine the final analysis result of the ship abnormal behavior. This overcomes the problem of completely relying on the prediction results for analysis when analyzing abnormal behavior, and avoids the influence of the deviation of the prediction results themselves on the analysis results.

[0045] Among them, when judging the initial analysis results, not only the prediction accuracy of the ship behavior prediction analysis model that obtains the prediction results is considered, but also the behavioral stability of the target ship is taken into account. The prediction results and the behavioral characteristics of the target ship itself are combined to further improve the accuracy of the ship abnormal behavior analysis.

[0046] The present application also provides a ship abnormal behavior analysis system, which executes the above-mentioned ship abnormal behavior analysis method, such as Figure 2 As shown, the system includes: an initial analysis result acquisition module, a ship behavior stability acquisition module, a model accuracy acquisition module and an output module; The initial analysis result acquisition module is used to obtain the ship data and ocean hydrological and meteorological environmental data of the target ship at the current moment, input the trained ship behavior prediction analysis model, and predict the behavior information of the target ship at the next moment; compare the prediction result with the ship behavior threshold to obtain the initial analysis result of the abnormal behavior of the ship; The ship behavior stability acquisition module obtains the behavior stability of the target ship by calculating the historical behavior information of the target ship; The model accuracy acquisition modulus is used to count the prediction accuracy of the ship behavior prediction analysis model at the current moment; The output module is used to judge the initial analysis results by combining the behavior stability of the target ship and the prediction accuracy of the ship behavior prediction analysis model, so as to determine the final analysis results of the abnormal behavior of the ship.

[0047] The present invention also provides an electronic device, including one or more processors and a memory.

[0048] The processor may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device to perform desired functions.

[0049] The memory may include one or more computer program products, and the computer program product may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory (cache), etc. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor may run the program instructions to implement a ship abnormal behavior analysis method and / or other desired functions of any embodiment of the present application described above. Various contents such as initial external parameters, thresholds, etc. may also be stored in the computer-readable storage medium.

[0050] In one example, the electronic device may further include: an input device and an output device, and these components are interconnected through a bus system and / or other forms of connection mechanisms (not shown). The input device may include, for example, a keyboard, a mouse, etc. The output device may output various information to the outside, including early warning information, braking force, etc. The output device may include, for example, a display, a speaker, a printer, a communication network and a remote output device connected thereto, etc.

[0051] Of course, for the sake of simplicity, components such as buses, input / output interfaces, etc. are omitted. In addition, the electronic device may further include any other appropriate components according to specific application conditions.

[0052] In addition to the above-mentioned methods and devices, an embodiment of the present application may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to implement the functions of a ship abnormal behavior analysis method provided by any embodiment of the present application.

[0053] The computer program product may be written in any combination of one or more programming languages ​​to write program codes for performing the operations of the embodiments of the present application, including object-oriented programming languages, such as Java, C++, etc., and conventional procedural programming languages, such as "C" language or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as an independent software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0054] In addition, an embodiment of the present application may also be a computer-readable storage medium on which computer program instructions are stored. When the computer program instructions are executed by a processor, the processor implements a method for analyzing abnormal ship behavior provided by any embodiment of the present application.

[0055] The computer readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can include, for example, but is not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0056] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the technical solutions of the embodiments of the present invention.

Claims

1. A method for analyzing abnormal behavior of a ship, characterized in that: The method comprises the following steps: S1: Obtain the current ship data and ocean hydrological and meteorological environmental data of the target ship, input the trained ship behavior prediction analysis model, and predict the behavior information of the target ship at the next moment; S2: Compare the prediction results in S1 with the ship behavior threshold to obtain the initial analysis results of the abnormal ship behavior; S3: The behavior stability of the target ship is obtained by calculating the historical behavior information of the target ship; S4: Calculate the prediction accuracy of the ship behavior prediction analysis model at the current moment; S5: Combine the behavior stability of the target ship and the prediction accuracy of the ship behavior prediction analysis model to judge the initial analysis results, so as to determine the final analysis results of the abnormal behavior of the ship.

2. A method for analyzing abnormal ship behavior according to claim 1, characterized in that: The ship behavior prediction and analysis model includes a GRU unit, a Spatial-Reduction Attention unit and a fully connected layer.

3. A method for analyzing abnormal ship behavior according to claim 1, characterized in that: The behavior information includes heading, speed, and longitude and latitude.

4. A method for analyzing abnormal ship behavior according to claim 3, characterized in that: In S3, the historical behavior information of the target ship is calculated to obtain the behavior stability of the target ship, which is specifically: S31: collecting historical behavior information of the target ship within a certain time range, and preprocessing the collected historical behavior information; S32: Calculate the heading, speed, and longitude and latitude in the pre-processed historical behavior information to obtain heading stability, speed stability, and longitude and latitude stability; S33: Based on the weighted method, the behavior stability of the target ship is obtained according to the heading stability, speed stability and longitude and latitude stability.

5. A method for analyzing abnormal ship behavior according to claim 1, characterized in that: In S5, the initial analysis results are judged in combination with the behavior stability of the target ship and the prediction accuracy of the ship behavior prediction analysis model, so as to determine the final analysis results of the abnormal behavior of the ship, specifically: Constructing a feature vector according to the behavior stability of the target ship and the prediction accuracy of the ship behavior prediction analysis model; According to the characteristic vector, the model is adjusted by analyzing the result to obtain the confidence of the initial analysis result; According to the confidence level and in combination with the initial analysis result, a final analysis result of the abnormal behavior of the ship is determined.

6. A method for analyzing abnormal ship behavior according to claim 5, characterized in that: According to the characteristic vector, the model is adjusted by analyzing the results to obtain the confidence of the initial analysis results, including: Normalizing the feature vector to obtain a normalized feature vector; According to the normalized feature vector, the model is adjusted by the trained analysis result to obtain a decision value corresponding to the initial analysis result; The decision value is converted to serve as the confidence level of the initial analysis result.

7. A method for analyzing abnormal ship behavior according to claim 6, characterized in that: The decision value is converted using an activation function.

8. A ship abnormal behavior analysis system, executing a ship abnormal behavior analysis method according to any one of claims 1 to 7, characterized in that: The system comprises: an initial analysis result acquisition module, a ship behavior stability acquisition module, a model accuracy acquisition module and an output module; The initial analysis result acquisition module is used to obtain the ship data and ocean hydrological and meteorological environmental data of the target ship at the current moment, input the trained ship behavior prediction analysis model, and predict the behavior information of the target ship at the next moment; compare the prediction result with the ship behavior threshold to obtain the initial analysis result of the abnormal behavior of the ship; The ship behavior stability acquisition module obtains the behavior stability of the target ship by calculating the historical behavior information of the target ship; The model accuracy acquisition modulus is used to count the prediction accuracy of the ship behavior prediction analysis model at the current moment; The output module is used to judge the initial analysis results by combining the behavior stability of the target ship and the prediction accuracy of the ship behavior prediction analysis model, so as to determine the final analysis results of the abnormal behavior of the ship.

Images