Illegal behavior full-scene feature recognition method and system in port intelligent monitoring

Abstract

The invention provides an illegal behavior full-scene feature recognition method and system in port intelligent monitoring, and relates to the technical field of intelligent monitoring. The method comprises the following steps: firstly, carrying out gridding division on a target port, processing operation information and meteorological data in a future first time interval, predicting violation behaviors with relatively high occurrence probability in the first time interval based on a machine learning model, then determining a simulation violation rule in advance based on the violation behaviors with the relatively high occurrence probability, and carrying out simulation on the violation rule according to the simulation violation rule. The method comprises the following steps: acquiring a plurality of sensing data, determining a first sensing data set, setting different weight information for sensing data with different importance degrees in the first sensing data set, finally predicting a plurality of pieces of first violation information based on the first sensing data, and verifying through a pre-determined violation simulation rule to determine target violation information. According to the technical scheme, multi-source data are analyzed and applied, and fine-grained prediction of illegal behaviors can be carried out in advance.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent monitoring, and particularly relates to a method and system for identifying full-scenario features of illegal behaviors in port intelligent monitoring. Background Art

[0002] In the technical field of port intelligent monitoring, although existing monitoring systems already have certain video analysis and anomaly detection capabilities, there are still many technical bottlenecks to be broken through. Currently, most systems rely on real-time video streams for the identification and alarm of illegal behaviors, lacking comprehensive consideration of future operation plans and environmental change factors, and it is difficult to achieve prior prediction and active prevention and control of potential illegal behaviors. Especially in the complex and changeable port environment, multi-source information such as meteorological conditions, ship dynamics, and loading and unloading plans has a significant impact on the occurrence of illegal events, and existing technologies have not effectively integrated this future time-series information for intelligent reasoning and risk prediction. Summary of the Invention

[0003] The purpose of the present invention is to provide a method and system for identifying full-scenario features of illegal behaviors in port intelligent monitoring, so as to solve the technical problem of inaccurate identification of port illegal behaviors in the prior art.

[0004] The present application proposes a method for identifying full-scenario features of illegal behaviors in port intelligent monitoring, and the method includes: S1: Preprocess the first operation information and the first meteorological information within the first time interval in the future to obtain the first multi-source operation information; S2: Input the first multi-source operation information into the first illegal information determination model to obtain the first illegal information; S3: Determine a number of first simulated illegal rules according to the first illegal information; S4: Determine the first sensor weight information according to the first illegal information, and determine the first sensing data set based on the first sensor weight information; S5: Input the first sensing data set into the first illegal behavior determination model to output a plurality of first illegal behaviors; S6: Based on the plurality of first illegal behaviors, determine a plurality of second simulated illegal rules from the number of first simulated illegal rules, and verify the plurality of first illegal behaviors through the plurality of second simulated illegal rules to obtain a plurality of target illegal behaviors.

[0005] Preferably, the S1 includes the following sub-steps: S11: Obtain the first operation information and the first meteorological information of the target port within the first time interval; S12: Perform first preprocessing and first spatio-temporal alignment processing on the first operation information and the first meteorological information to obtain first multi-source operation information.

[0006] Preferably, the first meteorological information includes: wind speed, wave height, visibility, precipitation probability, lightning warning data.

[0007] Preferably, the first preprocessing specifically includes: Convert the first operation information and the first meteorological information into a standardized JSON format with a unified timestamp; Resample the first meteorological information into a time series that matches the operation time granularity of the first operation information.

[0008] Preferably, S12 further includes: Generate a two-dimensional grid matrix based on the data after the first preprocessing and the first spatio-temporal alignment processing. In the two-dimensional grid matrix, each grid corresponds to at least one specified first operation information, and each grid also corresponds to a meteorological data information.

[0009] Preferably, the first violation information determination model is obtained through the following method: S21: Obtain multiple first historical sample data of the target port; S22: Perform standardization processing on each of the first historical sample data to obtain multiple first historical sample vectors; S23: Perform first clustering processing on the multiple first historical sample vectors to obtain multiple first clustering centers; S24: Obtain the first violation information determination model based on the multiple first clustering centers.

[0010] Preferably, S3 includes the following sub-steps: S31: Based on the first violation information, determine multiple first violation events; S32: Determine multiple first simulated violation rules according to the multiple first violation events and the first safe operation specifications.

[0011] Preferably, S4 includes the following sub-steps: S41: Determine multiple first high-probability violation information based on the first violation information; the first high-probability violation information includes first high-probability violation behaviors and first high-probability operation condition information; S42: For each of the first high-probability violation information, determine multiple first sensing data type information; S43: For each of the first high-probability operation condition information, determine multiple second sensing data type information according to the first sensor distribution information of the target port; S44: Determine first sensor weight information according to the multiple first sensing data type information and the multiple second sensing data type information; S45: Determine a first sensing data set based on the first sensor weight information and the multiple second sensing data type information.

[0012] Preferably, the S6 includes the following sub-steps: S61: For each of the first violation behaviors, determine multiple second simulated violation rules from the multiple first simulated violation rules. If the determination is successful, transfer to S62; otherwise, transfer to S63; S62: Verify the first violation behavior using the second simulated violation rule. When the verification fails, determine the first violation behavior as the target violation behavior; otherwise, output a misjudgment result and transfer to S64; S63: Perform a first violation determination on the first violation behavior. If the determination is successful, determine the first violation behavior as the target violation behavior; otherwise, output a misjudgment result and transfer to S64; S64: Obtain multiple target violation behaviors and output them.

[0013] The present application also proposes a system for identifying all-scenario features of violation behaviors in port intelligent monitoring, which is used to implement the method for identifying all-scenario features of violation behaviors in port intelligent monitoring described above.

[0014] The method and system for identifying all-scenario features of violation behaviors in port intelligent monitoring proposed in the present application relate to the technical field of intelligent monitoring. First, the target port is divided into grids, the operation information and meteorological data in the future first time interval are processed, and machine learning models are used to predict the violation behaviors with a relatively high probability of occurrence in the first time interval. Then, based on the violation behaviors with a relatively high probability of occurrence, simulated violation rules are determined in advance, and a first sensing data set is determined. Different weight information is set for sensing data with different importance levels in the first sensing data set. Finally, multiple first violation information is predicted based on the first sensing data and verified through the simulated violation rules determined in advance to determine the target violation information. The technical solution of the present invention analyzes and utilizes multi-source data, and can perform fine-grained prediction of violation behaviors in advance. Description of the Drawings

[0015] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only exemplary, and for those of ordinary skill in the art, without creative efforts, other implementation drawings can also be obtained based on the provided drawings.

[0016] Figure 1 It is the execution flowchart of the method for identifying all-scenario features of illegal behaviors in port intelligent monitoring in the present invention.

[0017] Figure 2 It is the schematic diagram of the grid coordinate system in the port operation area of the present invention. Specific embodiments

[0018] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0019] The following will detail the present invention in conjunction with the accompanying drawings and specific embodiments, where the illustrative embodiments and descriptions are only used to explain the present invention, but not to limit the present invention.

[0020] The following will detail the method and system for identifying all-scenario features of illegal behaviors in port intelligent monitoring of the present invention.

[0021] This embodiment proposes a method for identifying all-scenario features of illegal behaviors in port intelligent monitoring, and the specific process is as Figure 1 shown.

[0022] S1: Preprocess the first operation information and the first meteorological information within the first time interval in the future to obtain the first multi-source operation information.

[0023] In order to accurately predict the target illegal behaviors that are likely to occur within the first time interval in the future, in this step, the first operation information and the first meteorological information corresponding to the first time interval will be retrieved. Since the probability of violations generated by different types of operations under specified meteorological conditions will also vary, the illegal behaviors with a relatively high probability of occurrence within the first time interval can be preliminarily predicted through the target operation information and the target meteorological information within the first time interval.

[0024] The specific steps of S1 may include the following sub-steps: S11: Obtain the first operation information and the first meteorological information of the target port within the first time interval.

[0025] The target port is the port for which illegal behavior prediction is to be carried out. There are various types of operation forms within the area defined by the target port.

[0026] The first operation information includes various types of operation information. Specific operation information may include: ship berthing and unberthing operations, which refer to the operations carried out when a ship enters or leaves the port, including processes such as navigation, berthing, and unberthing; container operations, involving operations such as loading, unloading, stacking, unpacking, and re-packing of containerized goods; bulk dry cargo, general bulk cargo, and break-bulk cargo operations, which involve the loading, unloading, and storage of dry bulk goods such as coal, ore, and grain, as well as various break-bulk goods such as bagged, bundled, and barreled goods; crude oil and liquefied natural gas operations, specifically for the loading, unloading, and storage of liquid and gaseous goods such as petroleum, chemicals, and liquefied natural gas. The above operation types are examples and not exhaustive.

[0027] The first meteorological information may include various meteorological factors affecting port operations. Specifically, it may include refined forecast data such as wind speed, wave height, visibility, precipitation probability, and lightning warning.

[0028] The first operation information can obtain corresponding data by invoking the API interface of the port operation system, and the first meteorological information can obtain accurate meteorological data by accessing the meteorological forecast system.

[0029] S12: Perform the first preprocessing and the first spatio-temporal alignment processing on the first operation information and the first meteorological information to obtain the first multi-source operation information.

[0030] For a specified type of port operation information, the probability of its occurrence of illegal behavior will change significantly with the meteorological information. Therefore, in this step, first, it is necessary to perform the perfection processing of outliers on the two types of information through the first preprocessing operation, and achieve the unification of data types through formatting processing, and then perform spatio-temporal alignment processing on the two types of data to facilitate the subsequent correlation prediction of target illegal behaviors.

[0031] The first preprocessing specifically includes: Convert the first operation information and the first meteorological information into a standardized JSON format with a unified timestamp (UTC+8 time zone); Resample the first meteorological information into a time series that matches the operation time granularity of the first operation information (such as every hour / every 3 hours); The sliding window method is used to detect the missing time period of operation records (the window size is preferably 1 hour). The KNN algorithm is used to interpolate and complete the missing meteorological data, and obvious data anomaly points are corrected.

[0032] The first spatio-temporal alignment process refers to matching the spatio-temporal coordinates of the first operation information and the first meteorological information. This is because the coordinate positions of different first operation information are different, and the meteorological information at different positions is also different. For example, compared with the positions of berthing, onshore, and in the sea, meteorological information such as sea waves and visibility will have significant differences. The first spatio-temporal alignment process specifically includes: Establish a grid coordinate system for the port operation area (preferably a 500m×500m grid); Spatially associate the meteorological data with the position of the first operation information (such as quay No. 3 corresponding to meteorological grid G07, as Figure 2 shown).

[0033] According to the data after the first preprocessing and the first spatio-temporal alignment process, a two-dimensional grid matrix is generated. In the two-dimensional grid matrix, each grid corresponds to at least one specified first operation information, and each grid also corresponds to a meteorological data information.

[0034] In the case where one grid corresponds to multiple first operation information and multiple meteorological data, it is necessary to select the main operation information and the main meteorological data. The selection rules include: for operation information, select the first operation information with the largest area as the main operation information; for meteorological data, select several meteorological factors with the greatest influence relationship with the main operation information as the main meteorological data.

[0035] According to the two-dimensional grid matrix, the first multi-source operation information is determined. That is to say, in the first multi-source operation information, it contains the corresponding relationship between each first operation information and the corresponding meteorological data.

[0036] Specifically, in the first multi-source operation information, it includes the following four-dimensional data (time × position × operation type × meteorological information). For example, in the time period from 8:00 to 12:00 on the 10th, in the grid of (2, 2), the operation type generated is container unloading, and the meteorological information in this grid (such as wind force, rainfall, temperature change, wave height, and visibility conditions, etc.).

[0037] S2: Input the first multi-source operation information into the first violation information determination model to obtain the first violation information.

[0038] The first violation information determination model is used to determine the probabilities of various types of violation behaviors occurring within the first time interval. In the first violation information, several types of violation behavior information with relatively high occurrence probabilities within the first time interval are recorded, as well as the corresponding occurrence probabilities for each type of violation behavior information.

[0039] The first violation information determination model is obtained through the following steps: S21: Obtain multiple pieces of first historical sample data of the target port.

[0040] Each piece of the first historical sample data is operation data of the target port within a preset time period in the past. The operation data includes operation data composed of operation types, meteorological information, and violation information generated within a specified time interval and within a specified coordinate range.

[0041] S22: Perform standardization processing on each piece of the first historical sample data to obtain multiple first historical sample vectors.

[0042] Among them, in each of the first historical sample vectors, 5-dimensional data is included, including time, coordinates, operation type, meteorological information, and violation information.

[0043] S23: Perform a first clustering process on the multiple first historical sample vectors to obtain multiple first cluster centers.

[0044] Since each of the first historical sample vectors is in the form of a 5-dimensional vector, covering the violation information of the target port under different times, coordinate positions, operation information, and meteorological data, when the sample size is large enough, this is sufficient to represent the probabilities of violation behaviors occurring in different operation scenarios of the target port. For example, for the scenario of container unloading, in the multiple first historical sample vectors, it may cover the situations of violation information generated under different coordinates and different meteorological data conditions at the target port.

[0045] In order to further explore the occurrence patterns of violation behaviors, in this step, it is necessary to perform a first clustering process on the multiple first historical sample vectors. For example, the K-MEANS clustering algorithm can be used to obtain multiple cluster centers.

[0046] S24: Obtain the first violation information determination model based on the multiple first cluster centers.

[0047] That is to say, the first violation information determination model is mainly composed of the multiple first cluster centers.

[0048] Preferably, the first violation information determination model is also used to receive the input first multi-source operation information, and calculate the Euclidean distance between the first multi-source operation information and multiple first cluster centers, and use the first N first cluster centers with the shortest Euclidean distance as the target cluster center. The N can be selected according to the specific situation. For example, when it is selected as 3, three target cluster centers can be determined. Then, the violation behaviors corresponding to the three target cluster centers can be used as the first violation information corresponding to the first multi-source operation information. The probability of occurrence of each type of violation behavior is negatively correlated with the length of the corresponding Euclidean distance, that is, the shorter the Euclidean distance, the greater the probability of occurrence.

[0049] Preferably, the first violation information also includes all or part of the information in the first multi-source operation information corresponding to the high-probability violation, which may specifically include time, location coordinates, etc. The above information in the first violation information can be used to subsequently determine the corresponding sensor data.

[0050] S3: Determine a plurality of first simulated violation rules according to the first violation information.

[0051] Based on the first violation information, the boundary conditions corresponding to each type of violation can be determined. Specifically, stacking violations can be restricted through three-dimensional geographic fences, and operation process violations can be restricted through preset timing rules.

[0052] Through this step, multiple corresponding first simulated violation rules can be determined in advance based on the first violation information that has a high degree of fit with the first operation information and the first meteorological information in the first time period in the future. Therefore, when a certain type of violation is subsequently determined to have occurred, the corresponding simulated violation rules can be quickly retrieved to quickly verify the predicted violation, thereby ensuring the real-time and accuracy of the early warning.

[0053] The S3 specifically includes the following sub-steps: S31: Based on the first violation information, determine a plurality of first violation events.

[0054] In S2, multiple violations have been identified and the first violation information has been obtained. In this step, multiple first violation events can be determined from the first violation information. The first violation event refers to a violation that has a high probability of occurring within a first time interval in the future.

[0055] S32: Determine a plurality of first simulated violation rules according to the plurality of first violation events and the first safety operation specification.

[0056] Each of the first violation events is the literal description information of each violation behavior. The first safety operation specification is the regulation that must be followed to ensure safe operation. For example, when the wind speed ≥ 14 m / s, gantry crane operation is prohibited; when the visibility < 500 m, the tugboat dispatching frequency is reduced by 50%.

[0057] When specifically implementing this step, semantic analysis and similarity matching can be performed on the first violation event and the first safety operation specification, and the rule with the highest matching degree is used as the first simulated violation rule, so that corresponding multiple first simulated violation rules can be obtained for multiple first violation events.

[0058] S4: Determine the first sensor weight information according to the first violation information, and determine the first sensor data set based on the first sensor weight information.

[0059] Each type of the first violation information in S2 corresponds to different sensor data weights. For example, for the violation information of dangerous goods leakage, the relevant sensor data output by temperature and dangerous gas sensors is relatively important, while for the violation information of not wearing a safety helmet correctly, the relevant sensor data output by image sensors is usually relatively important. And when determining the weights of the sensor data, it is also necessary to extract the high-probability occurrence time and location of the corresponding violation event from the first violation information, so as to facilitate retrieving the sensor data at the corresponding time and location.

[0060] Therefore, through this step, the weight information corresponding to various types of sensor data can be determined based on the first violation information first, and the measured sensor data is weighted by the obtained first sensor weight information to obtain the first sensor set.

[0061] The S4 can specifically include the following sub-steps: S41: Determine multiple first high-probability violation information based on the first violation information.

[0062] In step S2, multiple violation behaviors with relatively high occurrence probabilities corresponding to the first multi-source operation information have been determined based on the first violation information determination model. Therefore, in this step, multiple high-probability violation behaviors and corresponding multiple operation condition information can be extracted from the first violation information.

[0063] Among them, the first high-probability violation information includes the first high-probability violation behavior and the first high-probability operation condition information. The first high-probability operation condition information includes the violation time information and the violation coordinate position information corresponding to each first high-probability violation behavior.

[0064] S42: For each of the first high-probability violation information, determine multiple first sensor data type information.

[0065] Since the characteristics of different violations vary, the types of sensing data used to detect such violations will also differ significantly. Therefore, in this step, for each of the first high-probability violation information, the types of sensing data that will be given key consideration when detecting such violation information are determined as the multiple first sensing data type information.

[0066] S43: For each of the first high-probability operating condition information, based on the first sensor distribution information of the target port, multiple second sensing data type information are determined.

[0067] In order to conduct all-round and multi-angle monitoring of the target port while taking into account the principles of economy and efficiency, various types of sensors are usually arranged in the target port according to certain layout specifications.

[0068] In this step, for each of the first high-probability operating condition information, multiple sensor nodes that will have a monitoring relationship with it in the sensor layout data of the target port are determined, so as to obtain the multiple second sensing data type information.

[0069] S44: Based on the multiple first sensing data type information and the multiple second sensing data type information, first sensor weight information is determined.

[0070] In this step, the number of sensors covered by the multiple second sensing data type information is greater than that of the multiple first sensing data type information. Therefore, in the multiple second sensing data type information, relatively high weights can be set for the multiple first sensing data type information, and relatively low weights can be set for the remaining multiple second sensing data type information.

[0071] Preferably, among the multiple first sensing data type information, weight differentiation can also be set according to human experience in a stepped pattern or other patterns, so as to make the determined first sensor weight information more in line with the actual situation.

[0072] S45: Based on the first sensor weight information and the multiple second sensing data type information, a first sensing data set is determined.

[0073] In this step, using the first sensor weight information, each of the multiple second sensing data type information is subjected to standardization and weighting processing to obtain the first sensing data set.

[0074] In the first sensing data set, all the sensing data types in the multiple second sensing data type information are included, but the weight of each sensing data type will vary due to the first sensor weight information.

[0075] S5: Input the first sensing data set into the first violation determination model to output multiple first violations.

[0076] Preferably, the first violation determination model is established based on the Multi-Granularity Feature Pyramid Network (MGFP-Net), and can analyze based on the current first sensing data set from three dimensions: macroscopic, mesoscopic, and microscopic, so as to accurately determine multiple first violations, where each first violation refers to a violation with a probability greater than a preset value.

[0077] The first violation determination model can comprehensively analyze and process the first sensing data set from three dimensions, so that the multiple first violations obtained are more accurate and in line with the actual situation.

[0078] In another embodiment, the first violation determination model can also be trained based on the ordinary Convolutional Neural Network model CNN, and the training method can adopt well-known means, which will not be elaborated here.

[0079] Among the multiple first violations determined, each includes violation type information, violation time information, and violation location information.

[0080] S6: Based on the multiple first violations, determine multiple second simulated violation rules from several first simulated violation rules, and verify the multiple first violations through the multiple second simulated violation rules to obtain multiple target violations.

[0081] Among the multiple first violations determined in step S5, there may be misjudgments. Therefore, it is necessary to find multiple first simulated violation rules corresponding to the multiple first violations from several first simulated violation rules determined in advance, and then for each first violation, verify it through the corresponding first simulated violation rule to determine whether each first violation is correctly judged, and determine the correctly judged multiple first violations as multiple target violations.

[0082] S6 includes the following sub-steps: S61: For each first violation, determine whether there is a corresponding first simulated violation rule. If so, go to S62; otherwise, go to S63.

[0083] In S3, multiple first simulated violation rules have been established in advance for high-probability violation information. Therefore, in this step, it is necessary to first determine whether there is a corresponding first simulated violation rule for each first violation.

[0084] S62: Verify the first violation using the first simulated violation rule. When the verification fails, determine the first violation as the target violation; otherwise, output a misjudgment result and proceed to S64.

[0085] Since the correct and compliant operating conditions are defined in the first simulated violation rule, when the verification fails, that is, when the first violation does not conform to the first simulated violation rule, it indicates that the first violation is indeed a violation, so it is determined as the target violation. Otherwise, it indicates that the first violation is a misjudgment.

[0086] S63: Conduct a first violation determination on the first violation. If the determination is successful, determine the first violation as the target violation; otherwise, output a misjudgment result and proceed to S64.

[0087] Preferably, the first violation determination can be performed based on expert experience or by setting determination rules.

[0088] S64: Obtain multiple target violations and output them.

[0089] This application also proposes a full-scenario feature recognition system for violations in port intelligent monitoring, which is used to execute the above-mentioned full-scenario feature recognition method for violations in port intelligent monitoring.

[0090] The full-scenario feature recognition method and system for violations in port intelligent monitoring proposed in this application relate to the technical field of intelligent monitoring. First, the target port is divided into grids, the operation information and meteorological data in the first future time interval are processed, and based on a machine learning model, the violations with a relatively high occurrence probability in the first time interval are predicted. Then, based on the violations with a relatively high occurrence probability, simulated violation rules are determined in advance, and a first set of sensing data is determined. Different weight information is set for the sensing data with different importance levels in the first set of sensing data. Finally, multiple first violation information is predicted based on the first set of sensing data and verified through the pre-determined simulated violation rules to determine the target violation information. The technical solution of the present invention analyzes and utilizes multi-source data and can perform fine-grained prediction of violations in advance.

[0091] The above is only a preferred embodiment of the present invention. Therefore, all equivalent changes or modifications made according to the structure, features, and principles described in the scope of this invention patent application are included in the scope of this invention patent application.

Claims

1. A method for identifying the full-scenario characteristics of illegal behaviors in port intelligent monitoring, characterized in that, The method includes: S1: Preprocess the first operation information and the first meteorological information within the first future time interval to obtain the first multi-source operation information; The S1 includes the following sub-steps: S11: Obtain the first operation information and the first meteorological information of the target port within the first time interval; S12: Perform the first preprocessing and the first spatio-temporal alignment processing on the first operation information and the first meteorological information to obtain the first multi-source operation information; The first preprocessing specifically includes: Convert the first operation information and the first meteorological information into a standardized JSON format with a unified timestamp; Resample the first meteorological information into a time series that matches the operation time granularity of the first operation information; S2: Input the first multi-source operation information into the first violation information determination model to obtain the first violation information; S3: Determine a number of first simulated violation rules according to the first violation information; S4: Determine the first sensor weight information according to the first violation information, and determine the first sensing data set based on the first sensor weight information; S5: Input the first sensing data set into the first violation behavior determination model to output multiple first violation behaviors; S6: Based on multiple first violation behaviors, determine multiple second simulated violation rules from a number of first simulated violation rules, and verify the multiple first violation behaviors through the multiple second simulated violation rules to obtain multiple target violation behaviors.

2. The method for identifying full-scenario features of violations in port intelligent monitoring according to claim 1, wherein, The first meteorological information includes: wind speed, wave height, visibility, precipitation probability, lightning warning data.

3. The method for identifying all-scenario features of illegal behaviors in port intelligent monitoring according to claim 2, wherein The S12 further includes: Generate a two-dimensional grid matrix according to the data after the first preprocessing and the first spatio-temporal alignment processing. In the two-dimensional grid matrix, each grid corresponds to at least one specified first operation information, and each grid also corresponds to a meteorological data information.

4. The method for identifying all-scenario features of illegal behaviors in port intelligent monitoring according to claim 3, wherein The first violation information determination model is obtained in the following manner: S21: Obtain multiple first historical sample data of the target port; S22: Perform standardization processing on each first historical sample data to obtain multiple first historical sample vectors; S23: Perform the first clustering processing on multiple first historical sample vectors to obtain multiple first clustering centers; S24: Obtain the first violation information determination model according to multiple first clustering centers.

5. The method for identifying all-scenario features of illegal behaviors in port intelligent monitoring according to claim 4, wherein The S3 includes the following sub-steps: S31: Determine multiple first violation events based on the first violation information; S32: Determine multiple first simulated violation rules according to multiple first violation events and the first safety operation specification.

6. The method for identifying the full-scenario features of illegal behaviors in port intelligent monitoring according to claim 5, wherein The S4 includes the following sub-steps: S41: Determine multiple first high-probability violation information based on the first violation information; the first high-probability violation information includes the first high-probability violation behavior and the first high-probability operation condition information; S42: For each first high-probability violation information, determine multiple first sensing data type information; S43: For each of the first high-probability operation condition information, determine a plurality of second sensing data type information according to the first sensor distribution information of the target port; S44: Determine first sensor weight information according to the plurality of first sensing data type information and the plurality of second sensing data type information; S45: Determine a first sensing data set based on the first sensor weight information and the plurality of second sensing data type information.

7. The method for identifying the full-scenario features of illegal behaviors in port intelligent monitoring according to claim 6, wherein The S6 includes the following sub-steps: S61: For each of the first violation behaviors, determine a plurality of second simulation violation rules from the plurality of first simulation violation rules. If the determination is successful, transfer to S62; otherwise, transfer to S63; S62: Verify the first violation behavior using the second simulation violation rule. When the verification fails, determine the first violation behavior as the target violation behavior; otherwise, output a misjudgment result and transfer to S64; S63: Perform a first violation determination on the first violation behavior. If the determination is successful, determine the first violation behavior as the target violation behavior; otherwise, output a misjudgment result and transfer to S64; S64: Obtain and output a plurality of target violation behaviors.

8. A system for identifying all-scenario characteristics of violation behaviors in port intelligent monitoring, which is used to implement the method for identifying all-scenario characteristics of violation behaviors in port intelligent monitoring according to any one of claims 1-7.

Images