A method and system for identifying false declaration of dangerous goods container export declaration

Abstract

The application provides a method and system for identifying dangerous goods container export customs declaration false reporting behavior, and relates to the technical field of dangerous goods transportation supervision. The method comprises the following steps: establishing a high-risk point information database and setting an electronic fence, and configuring multiple trigger rules; connecting with customs, traffic and terminal EIR data, and screening containers declared as ordinary goods as the supervision object; obtaining the track of the vehicle from emptying the container to returning the container to the weight, and using a multi-dimensional risk assessment model containing space, time, speed, historical risk items and risk credibility correction factors to determine whether the vehicle effectively enters the electronic fence; if it enters and is declared as ordinary goods, an alarm is generated and pushed to law enforcement personnel. The system comprises corresponding modules, and dynamically updates the point database and model parameters based on law enforcement feedback. The application significantly reduces false positives, and forms a self-feedback cycle of supervision-alarm-inspection-update.

Description

Technical Field

[0001] This invention relates to the field of dangerous goods transportation supervision technology, specifically to a method and system for identifying false declarations in customs declarations for dangerous goods containers exported based on multi-source data fusion and multi-dimensional coupled risk assessment. Background Technology

[0002] In the export transportation of dangerous goods containers, some shippers, in order to circumvent the strict regulatory requirements for dangerous goods transportation and reduce transportation costs, engage in illegal practices such as falsely declaring dangerous goods as ordinary goods for export customs declaration. This kind of false declaration not only violates relevant customs and logistics supervision laws and regulations, but also poses a significant threat to safety accidents such as fires, explosions, and leaks of toxic and hazardous substances due to improper transportation and handling of dangerous goods. It also severely threatens the personal safety, property safety, and ecological environment of port yards and surrounding transportation routes. Therefore, how to accurately and efficiently identify false declarations in the export customs declaration of dangerous goods containers has become a crucial technical problem that customs and port regulatory authorities urgently need to solve.

[0003] Currently, regulatory measures against false declarations of dangerous goods containers mainly rely on manual spot checks, document verification, and post-event traceability. These methods suffer from low overall efficiency, limited coverage, and significant delays in detection. Manual inspections are limited by the inspection ratio and cannot achieve full coverage of the massive number of export containers, resulting in a persistently high rate of missed detections of false declarations. Traditional regulatory models lack dynamic monitoring of the entire container transportation process, making it difficult to detect anomalies in advance during loading and transportation. Often, traceability and verification can only be carried out after a safety incident, failing to achieve pre-emptive warnings and effective interception during the process. Furthermore, various regulatory data, such as customs declaration data, transportation data, and terminal operation data, are independent of each other, and a multi-source data collaborative analysis mechanism has not yet been established, further exacerbating the situation of delayed detection and passive handling of false declarations.

[0004] To alleviate the aforementioned problems, the industry has gradually adopted technologies such as vehicle trajectory monitoring, electronic fences, and abnormal stop detection for transportation anomaly identification. For example, CN114357094B discloses a method for monitoring hazardous chemical transportation based on vehicle trajectory and geofencing, which identifies anomalies by judging the vehicle's entry into a specific area and the duration of its stay; CN115758095A discloses a multi-dimensional feature anomaly integral model, which constructs an anomaly integral judgment mechanism from spatial, temporal, and speed aspects. However, existing technologies still suffer from common and insurmountable defects: First, anomaly detection relies heavily on single thresholds or simple rule triggers, judging solely based on independent conditions such as entering an area, exceeding the permitted time limit, and low speed. Legitimate behaviors such as normal vehicle passage, traffic congestion, and waiting for traffic signals are easily misjudged as abnormal, leading to a persistently high false alarm rate. Second, existing technologies do not construct composite behavioral models targeting the typical operational characteristics of dangerous goods container deception. Dangerous goods deception typically presents a continuous composite behavior of "lifting empty containers, entering high-risk locations, prolonged stays, low-speed loading and unloading, and returning loaded containers." Existing solutions only simply weight or superimpose abnormal features in terms of space, time, and speed, failing to identify true deception patterns involving multiple coupled features, thus limiting accuracy. Third, the parameters of existing risk assessment models are mostly fixed, lacking adaptive optimization capabilities. High-risk areas and judgment rules cannot be dynamically iterated based on enforcement inspection results, leading to a continuous decline in recognition accuracy after long-term operation, making it difficult to cope with constantly evolving deception methods. Fourth, existing technologies do not integrate with customs declaration information and terminal equipment interchange forms. The lack of in-depth linkage between Receipt (EIR) information and key vehicles that declare ordinary goods but enter high-risk areas for dangerous goods cannot accurately identify key vehicles under supervision, resulting in insufficient targeted supervision and serious waste of resources.

[0005] In summary, existing regulatory methods are essentially extensions and improvements of traditional trajectory monitoring and rule-based early warning systems, rather than specialized, accurate, low-false-report, and iterative intelligent regulatory methods designed for scenarios involving false declarations in dangerous goods container export customs declarations. The industry urgently needs a novel technological solution for identifying false declarations of dangerous goods. Summary of the Invention

[0006] The technical problem this invention aims to solve is: addressing the issues of existing methods for identifying false declarations in dangerous goods container export customs declarations, such as high reliance on manual labor, limited coverage, high false alarm rate, inability to identify complex false declaration patterns, fixed model parameters that cannot be adaptively optimized, and ineffective integration of multi-source regulatory data. This invention provides a fully intelligent identification method and system that can deeply integrate multi-source data from customs, transportation, and terminals, construct a multi-dimensional coupled risk assessment model for complex dangerous goods loading behaviors, introduce a risk credibility correction mechanism to suppress false alarms, and support dynamic adaptive updates based on law enforcement feedback. This aims to achieve the regulatory goals of pre-emptive warning, precise interception during the process, and post-event optimization.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A method for identifying false declarations in customs declarations for the export of dangerous goods containers includes the following steps: Establish a database of high-risk locations for false declarations in customs declarations of dangerous goods containers exported; Electronic fences are set up based on high-risk geographical locations in the high-risk location information database to complete the control of the regulatory area, and multiple trigger rules are configured. The multiple trigger rules include risk judgment conditions in four aspects: space, time, speed and historical entry frequency. The multiple trigger rules adopt a comprehensive risk assessment model that includes a risk credibility correction factor. The risk credibility correction factor is used to jointly correct spatial risk, time risk and speed risk to suppress false high-risk scores caused by a single risk factor. By connecting with customs export information data, transportation information on container vehicles, and terminal yard equipment handover information, based on the obtained information such as container cargo type, container vehicle license plate information, and container vehicle pick-up and return times in and out of the yard, container vehicles with export shipping information as general cargo are screened out, and a list of container vehicles that need to be supervised is formed. The driving trajectory of the monitored vehicle during the period from the empty container leaving the site to the loaded container entering the site is obtained, and the comprehensive risk assessment model including the risk credibility correction factor is used to determine whether the monitored vehicle has effectively entered the electronic fence range. If the monitored vehicle effectively enters the electronic fence area and its export consignment information indicates it is ordinary cargo, an alarm signal will be generated and pushed to the law enforcement personnel's terminal, prompting them to intercept and inspect the suspected illegal container.

[0008] On the other hand, the present invention also provides a system for identifying false declarations in customs declarations for the export of dangerous goods containers, comprising: The High-Risk Location Information Database Establishment Module is used to establish a high-risk location information database for false declarations of dangerous goods container export customs declarations. The electronic fence deployment module is used to set up electronic fences based on high-risk geographical locations in the high-risk location information database, complete the deployment of the monitored area, and configure multiple trigger rules. The multiple trigger rules include risk judgment conditions in four aspects: space, time, speed, and historical entry count. The multiple trigger rules adopt a comprehensive risk assessment model that includes a risk credibility correction factor. The risk credibility correction factor is used to jointly correct the spatial risk, time risk, and speed risk to suppress false high-risk scores caused by a single risk factor. The regulatory list generation module is used to connect with customs department's cargo export information data, transportation department's container vehicle information, and terminal yard equipment handover form information. Based on the container cargo type, container vehicle license plate information, and container vehicle pick-up and return time in and out of the yard in the obtained information, it screens out container vehicles whose export consignment information is general cargo and forms a list of container vehicles that need to be regulated. The trajectory acquisition and determination module is used to acquire the driving trajectory of the monitored vehicle from the time the empty container leaves the site to the time the loaded container enters the site, and to use the comprehensive risk assessment model containing the risk credibility correction factor to determine whether the monitored vehicle has effectively entered the electronic fence range. The alarm generation and push module is used to generate an alarm signal and push it to the law enforcement personnel's terminal when the monitored vehicle effectively enters the electronic fence range and its export consignment information is ordinary goods. The law enforcement feedback and dynamic update module is used to receive the on-site inspection results from law enforcement officers and to dynamically update and optimize the high-risk location information database and risk assessment model parameters based on the inspection results.

[0009] Compared with the prior art, the beneficial effects of the present invention are: (1) By connecting with customs declaration data, transportation vehicle data and terminal EIR data, container vehicles declared as general cargo are automatically screened and a regulatory list is formed, which clarifies the regulatory objects and regulatory time intervals, avoids redundancy in the regulatory scope, and achieves accurate locking of vehicles suspected of false declarations, thus solving the problem of existing multi-source data being independent of each other and insufficient regulatory targeting.

[0010] (2) By constructing a multi-trigger judgment mechanism that includes a database of high-risk locations and electronic fences, high-risk areas for false alarms are identified from a spatial perspective. Effective entry judgment is made based on multiple risk thresholds, including comprehensive space, time, speed and historical entry counts, eliminating invalid triggering situations such as normal vehicle passage and traffic congestion, which significantly reduces the false alarm rate.

[0011] (3) In the process of comprehensive risk assessment, a risk credibility correction factor is introduced to jointly characterize spatial risk, time risk and speed risk. The risk credibility is significantly improved only when the three behavioral characteristics appear simultaneously and form a dangerous goods loading behavior pattern. This effectively suppresses false high-risk scores caused by abnormal fluctuations of a single risk factor, solves the problem of false alarms caused by simple rule triggering in the existing technology, and greatly improves the accuracy of identifying real dangerous goods loading behavior.

[0012] (4) By receiving the on-site inspection results from law enforcement feedback, the system automatically extracts new high-risk locations to achieve dynamic expansion of the information database. Based on the statistical analysis of valid alarm and false alarm samples, the system dynamically adjusts the weight of each risk item, alarm threshold, and risk credibility adjustment parameters, so that the risk assessment model can be continuously iterated and optimized with the accumulation of law enforcement data. This solves the problem of fixed parameters in the existing model and decreased recognition accuracy after long-term operation, forming a self-feedback loop of sequential connection between supervision, alarm, inspection and update, continuously improving recognition accuracy, and realizing full-process intelligent supervision of pre-warning, in-process precise interception and post-event optimization.

[0013] (5) By integrating the functional modules corresponding to each step of the above method into a complete identification system, the entire process from data collection, list generation, trajectory determination, alarm push to dynamic update is automated. The modules work together and the data is stored in full, which further improves the efficiency of supervision and the reliability of system operation.

[0014] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating a method for identifying false declarations in customs declarations for the export of dangerous goods containers, according to an embodiment of the present invention. Figure 2 This is a structural block diagram of a system for identifying false declarations in customs declarations for the export of dangerous goods containers, provided in an embodiment of the present invention. Figure 3 This is a schematic diagram illustrating the verification and comparison of risk credibility correction factors in one embodiment of the present invention. Detailed Implementation

[0016] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the following embodiments are only used to explain the present invention and are not intended to limit the scope of protection of the present invention. Those skilled in the art can adapt and adjust the preset parameters and docking systems in the embodiments according to actual regulatory requirements.

[0017] Example 1 like Figure 1As shown in the figure, this embodiment provides a method for identifying false declarations in customs declarations for the export of dangerous goods containers, which specifically includes the following steps S10 to S60.

[0018] S10: Establish a database of high-risk locations for false declarations in the export customs declaration of dangerous goods containers.

[0019] First, high-risk locations are screened. Based on customs records of historical misdeclarations of dangerous goods, safety risk assessments of hazardous chemical production and storage enterprises, and multi-dimensional information such as the risk level of dangerous goods operations at ports and site supervision levels, high-risk geographical locations related to misdeclarations of dangerous goods container export declarations are screened. These high-risk geographical locations specifically include: dangerous goods factories with a long history of dangerous goods operations and a history of misdeclarations; dangerous goods storage yards that do not implement dangerous goods control measures according to regulations; and temporary locations with frequent dangerous goods loading and unloading operations, among other high-risk locations for dangerous goods operations. The specific screening criteria based on historical violation records and the safety risk level of hazardous chemical production and storage enterprises are: one or more violations within the past 5 years; and a safety risk level of red (below 60 points), orange (below 60 to 75 points), or yellow (below 75 to 90 points).

[0020] Secondly, geographic information collection is conducted. Precise geographic coordinates, including longitude and latitude, are collected for each of the aforementioned high-risk geographical locations through satellite positioning, on-site surveying, and other methods. Combined with the actual layout of the site, the physical boundaries of the work areas, entrances and exits, loading and unloading areas, and supporting storage areas covered by each location are determined.

[0021] Next, the information database is constructed and configured. Based on the collected geographic coordinate information and physical boundary range, a structured high-risk location information database is built in the monitoring system. Each record in this database contains at least the following information: location name, geographic coordinates, physical boundary range, risk level, and main operation category. The database is also configured with dynamic update functionality to support operations such as adding locations, modifying information, and adjusting risk levels.

[0022] S20: Set up electronic fences based on high-risk locations to complete the control of the regulatory area.

[0023] First, data retrieval is performed. The monitoring system automatically retrieves the geographic coordinates and physical boundaries of all high-risk geographic locations from the high-risk location information database constructed in step S10.

[0024] Secondly, electronic fences are drawn. An electronic fence is independently drawn for each high-risk geographical location on the electronic map of the monitoring system. The boundaries of the drawn electronic fences perfectly match the physical boundaries of the corresponding locations, ensuring coverage of the work area, entrances / exits, and supporting areas of that location, thus achieving precise delineation of high-risk areas.

[0025] Next, configure the vehicle triggering rules for the geofence. These rules are used to determine whether a vehicle has effectively entered the geofence area, and specifically include setting coordinate determination rules and comprehensive risk threshold determination rules.

[0026] Among them, the coordinate determination rule is used to determine whether the real-time positioning coordinates of the vehicle fall within the physical boundary of the electronic fence, as a preliminary triggering condition.

[0027] The comprehensive risk threshold determination rule is as follows: after a vehicle meets the coordinate triggering condition, its dwelling risk value within the electronic fence is further calculated. Only when this dwelling risk value is not lower than the preset risk threshold is it determined to have effectively entered the electronic fence range. The calculation of the dwelling risk value adopts the following multi-dimensional coupled risk assessment model.

[0028] Define the vehicle's comprehensive risk value as The calculation formula is: In the formula, , , , These are spatial risk weights, temporal risk weights, velocity risk weights, and historical behavior risk weights, each satisfying a normalization condition: ,and . This is a risk credibility correction factor.

[0029] Space risk items Defined as: In the formula, T represents the total time the vehicle spends within the electronic fence. and Let represent the distances between the vehicle and the centers of the first and second high-risk areas at time t, respectively. , represents the scale parameter for the corresponding region. This formula indicates that the closer a vehicle is to the center of a high-risk area, the higher the spatial risk, and that risks from multiple regions can overlap.

[0030] Time risk items Defined as: In the formula, This is a time scale constant. This formula is used to describe the characteristic that the longer a vehicle remains stationary, the greater the suspicion of an anomaly.

[0031] Speed ​​risk item Defined as: In the formula, v(t) is the instantaneous speed of the vehicle at time t. This is a reference speed constant. This formula indicates that the speed risk value is higher when the vehicle is running at low speed or stationary.

[0032] Historical risk items Defined as: In the formula, N represents the total number of times the vehicle has entered high-risk areas in the past. This formula is used to characterize the cumulative impact of historical abnormal behavior.

[0033] Risk credibility correction factor Defined as: In the formula, This is a dynamic credibility adjustment parameter, whose value is periodically updated based on historical law enforcement feedback results. This adjustment factor is constructed... The spatial risk, temporal risk, and velocity risk are represented by a joint product: only when the above three behavioral characteristics occur simultaneously and all remain at a high level, When the value approaches 1, the risk credibility increases significantly; when only a single anomalous risk characteristic exists (e.g., only spatial risk is high while temporal and velocity risks are very low), The value is kept at a low level, thereby suppressing false high-risk scores caused by a single risk factor and improving the reliability of risk identification results.

[0034] Set preset risk threshold .when At that time, it was determined that the target vehicle had a high risk of illegal loading and an early warning was triggered.

[0035] In this embodiment, as an example and not a limitation, the initial parameters of the above risk assessment model are set as follows: Spatial risk weight Time risk weight Speed ​​risk weight Historical behavior risk weights Time scale constant minutes; reference velocity constant km / h; regional scale parameters Meters; Preset risk threshold Dynamic credibility adjustment parameters The above parameters are set during the initial system runtime and are subsequently dynamically and adaptively adjusted based on the law enforcement feedback results from step S60.

[0036] Finally, data synchronization is performed. The location information, physical boundary range, and configured triggering rules of all drawn electronic fences are synchronized to the system's trajectory acquisition and judgment module to complete the full-process control of the monitored area.

[0037] S30: Connect to multi-source data and screen to form a list of vehicles under supervision.

[0038] First, information integration is implemented. The regulatory system integrates in real time with the customs department's cargo export information system, the transportation department's container vehicle information system, and the terminal yard's EIR information system to automatically obtain the following information: cargo consignment declaration type (general cargo or dangerous goods), container number, carrier vehicle license plate number, empty container pickup time, and loaded container return time.

[0039] Secondly, information cleaning and screening are performed. The system automatically cleans and processes the acquired information, filtering out container vehicle records whose cargo consignment declaration type is clearly marked as dangerous goods, and retaining only container vehicle records whose declaration type is ordinary goods, as the key targets for investigating false declarations.

[0040] Next, the monitoring time interval is marked. Based on the vehicle's license plate information, the system automatically links the empty container pickup and drop-off records and the loaded container drop-off records of the same vehicle, marking the time period from the vehicle's empty container pickup departure to its loaded container drop-off as the vehicle's monitoring time interval. Subsequent tracking and judgment will only be performed on the vehicle's travel trajectory within this monitoring time interval.

[0041] Next, the regulatory list is generated and synchronized. Based on the above screening results and the marked regulatory time intervals, the system generates a structured list of container vehicles requiring regulation. Each record in this list includes at least: license plate number, container number, declared cargo type, and regulatory time interval (time of empty container pickup and time of loaded container return). The system synchronizes this list to the trajectory acquisition and judgment module in real time.

[0042] S40: Obtain the driving trajectory of the monitored vehicle and determine whether it has effectively entered the electronic fence.

[0043] First, trajectory data is acquired. The system interfaces with the satellite positioning system (such as GPS, BeiDou, etc.) of the monitored vehicles. Based on the license plate information in the monitored vehicle list generated in step S30, it achieves accurate matching of vehicle positioning data and automatically acquires the real-time driving trajectory data of the corresponding vehicle within its monitored time period. The trajectory data includes the vehicle's real-time positioning coordinates (longitude and latitude) and driving timestamp, with a positioning accuracy of no less than 10 meters.

[0044] Secondly, the trajectory is compared with the electronic fence. The system automatically compares the real-time driving trajectory data of the acquired vehicle with the location information, physical boundary range and configured trigger judgment rules of all electronic fences deployed in step S20, and checks whether the vehicle positioning coordinates corresponding to each timestamp meet the coordinate triggering conditions of the electronic fence.

[0045] Next, a valid entry determination is performed. Based on the comparison results and the comprehensive risk threshold determination rule set in step S20, it is determined whether the monitored vehicle has effectively entered the electronic fence area. Specifically, when the real-time positioning coordinates of the vehicle fall within the physical boundary of a certain electronic fence, the coordinate triggering condition is met, and the system then uses the above formula to calculate the comprehensive risk value of the vehicle within that electronic fence. .when When a vehicle enters the electronic fence area, the system determines that it has effectively entered the area. If the entry is deemed valid, the system automatically records the vehicle's entry time, the duration of continuous stay within the fence, the name and geographic coordinates of the corresponding high-risk location, and synchronizes the determination result to the alarm push and recording module.

[0046] S50: Generates an alarm signal and pushes it to the law enforcement personnel's terminal.

[0047] First, an alarm signal is generated. When the system determines that a monitored vehicle has effectively entered the electronic fence area, and the export shipment declaration type of the container carried by the vehicle is general cargo, an alarm signal for suspected false declaration of dangerous goods is automatically generated. The alarm signal contains complete key information, specifically: the vehicle's license plate number, the vehicle's monitoring time range, the time of entry into the electronic fence, the duration of continuous stay, the name and geographical coordinates of the corresponding high-risk geographical location, the container number, an EIR information screenshot, and the vehicle's current real-time location.

[0048] Secondly, alarm signals are pushed out. The system pushes the generated alarm signals in real time to the mobile terminal devices (such as mobile phones and tablets) of designated law enforcement personnel and the monitoring back-end terminal through wired networks or 5G / 4G wireless networks, ensuring that law enforcement personnel can obtain complete information on suspected illegal activities as soon as possible.

[0049] Simultaneously, alarm records are retained. While sending alarm signals, the system automatically retains complete alarm records in the monitoring backend. These records include at least: alarm generation time, alarm details, all information about the involved vehicle and container, the status of receipt by law enforcement personnel, and the status of subsequent handling, enabling full traceability of alarm information and providing evidentiary support for subsequent violation investigations and administrative penalties.

[0050] S60: Risk locations and model parameters are dynamically updated based on law enforcement feedback.

[0051] First, the enforcement results are transmitted back. Enforcement officers conduct on-site inspections of the target vehicles based on system alerts and upload the inspection results to the monitoring platform via terminal devices. The inspection results must include at least one of the following three categories: confirmation of false declaration of dangerous goods transportation; no false declaration of dangerous goods found; or other abnormal transportation behavior discovered. Simultaneously, the system records the corresponding vehicle's complete trajectory information, stopping locations, stopping time, and risk score results.

[0052] Secondly, the high-risk location database is dynamically updated. When law enforcement confirms a false declaration of hazardous materials, the system automatically extracts the coordinates and trajectory characteristics of the vehicle's actual loading location to determine if the location has been included in the existing high-risk location database. If the location is not yet included, the system automatically creates a high-risk location record for it, collects its geographical coordinates and physical boundary range, and adds it to the high-risk location database, thus achieving dynamic expansion of the database.

[0053] Next, the risk parameters are adaptively updated. Based on the law enforcement feedback, the system counts the number of valid alarm samples and the number of false alarm samples. For false alarm samples, the system analyzes their spatial risk value. Time risk value Speed ​​risk value and historical behavioral risk values The system analyzes the distribution characteristics of each risk item; for valid alarm samples confirming violations, it analyzes their typical loading behavior characteristics. Based on the above statistical analysis results, the system dynamically adjusts the weights of each risk item. , , , and risk alarm thresholds This allows the risk assessment model to gradually adapt to the actual regulatory environment and continuously improve the accuracy of identification.

[0054] Next, the risk credibility parameters are optimized. This involves adjusting the credibility parameters in step S20 above. The system dynamically adjusts based on law enforcement feedback: when the false alarm rate increases, the adjustment is appropriately increased. The value is adjusted to reduce the system's sensitivity to local anomalies, thereby reducing false alarms; when the proportion of missed alarms increases, the value is appropriately reduced. This enhances the system's ability to identify patterns in hazardous materials loading, thereby reducing false negatives. Through continuous optimization... The parameters ensure that the risk credibility correction model always maintains superior recognition performance.

[0055] Finally, the updated model was put back into operation. After updating the risk points and optimizing the model parameters, the system generated a new set of risk assessment model parameters and applied it to the risk identification process of subsequent monitored vehicles. As law enforcement feedback data continues to accumulate, the risk model is continuously iterated and optimized, gradually improving the accuracy of identifying false declarations in the export customs declaration of dangerous goods containers, while simultaneously reducing the false alarm rate and the missed alarm rate. This forms a self-feedback loop in which supervision, alarm, inspection, and updating are sequentially linked, and the system continuously improves its identification performance as law enforcement data accumulates.

[0056] Example 2 like Figure 2 As shown, this embodiment provides a system for identifying false declarations in customs declarations for the export of dangerous goods containers. This system is used to execute the method described in Embodiment 1 above. Specifically, the system includes: a high-risk location information database establishment module 100, an electronic fence deployment module 200, a regulatory list generation module 300, a trajectory acquisition and judgment module 400, an alarm generation and push module 500, an enforcement feedback and dynamic update module 600, and a data storage module for storing high-risk location information, electronic fence configuration information, a list of regulatory vehicles, trajectory data, alarm records, and enforcement feedback results. Figure 2 (Not shown in the diagram; can be integrated into individual modules or set up independently). Modules interact with each other via wired or wireless communication networks to collaboratively complete the entire process of identifying and monitoring false reporting behavior.

[0057] The High-Risk Location Information Database Establishment Module 100 is used to establish a database of high-risk locations for false declarations of dangerous goods container exports. This module specifically performs the following operations: Based on historical customs records of false declarations of dangerous goods, safety risk level assessments of dangerous chemical production and storage enterprises, and multi-dimensional information such as the risk level of dangerous goods handling categories and site supervision levels at ports, it identifies high-risk geographical locations; it collects precise geographic coordinates of each high-risk location through satellite positioning and on-site mapping, and determines the physical boundaries of the operating areas, entrances / exits, loading / unloading areas, and supporting storage areas covered by each location, combined with the actual site layout; it constructs a structured high-risk location information database based on the collected geographic coordinates and physical boundaries, and configures a dynamic update function for the database. This module sends the completed high-risk location information database to the electronic fence deployment module 200 and the law enforcement feedback and dynamic update module 600.

[0058] The electronic fence deployment module 200, connected to the high-risk location information database establishment module 100, is used to set up electronic fences based on high-risk geographical locations in the database, completing the deployment of the monitored area and configuring multiple triggering rules based on time, space, speed, and historical entry counts. Specifically, this module performs the following operations: retrieves the geographic coordinates and physical boundary ranges of all high-risk geographical locations in the database; independently draws an electronic fence for each high-risk geographical location on the electronic map of the monitoring system, ensuring the boundaries of the electronic fence perfectly match the physical boundary range of the location; configures the vehicle triggering judgment rules for the electronic fence, including coordinate judgment rules and comprehensive risk threshold judgment rules, where the comprehensive risk threshold judgment rule adopts the multi-dimensional coupled risk assessment model defined in Example 1; and synchronizes the location information, physical boundary ranges, and triggering judgment rules of all electronic fences to the trajectory acquisition and judgment module 400. The electronic fence configuration data generated by the electronic fence deployment module 200 is simultaneously stored in the data storage module.

[0059] The regulatory list generation module 300 is used to connect with customs export information data, transportation container vehicle information, and terminal yard equipment handover document information to screen and generate a list of container vehicles requiring supervision. This module specifically performs the following operations: It connects in real-time with the customs export information system, transportation container vehicle information system, and terminal yard EIR information system to automatically obtain cargo consignment declaration type, container number, vehicle license plate number, empty container pickup time, and loaded container return time; it automatically cleans the information, filtering out vehicle records whose declaration type is clearly dangerous goods, retaining only vehicle records declared as general cargo; it associates the empty container pickup and loaded container return records of the same vehicle, marking the supervision time interval; it generates a structured list of container vehicles requiring supervision and synchronizes this list in real-time to the trajectory acquisition and judgment module 400. The list data generated by the regulatory list generation module 300 is also stored in the data storage module.

[0060] The trajectory acquisition and judgment module 400 is connected to the electronic fence deployment module 200 and the supervision list generation module 300, respectively. It is used to acquire the driving trajectory of the supervised vehicle during the period from the removal of an empty container to the return of a loaded container, and to determine whether the supervised vehicle has effectively entered the electronic fence area. Specifically, this module performs the following operations: It interfaces with the satellite positioning system of the supervised vehicle, accurately acquires the real-time driving trajectory data of the corresponding vehicle within the supervision time interval based on the license plate information in the supervised vehicle list. The trajectory data includes real-time positioning coordinates and a driving timestamp; it automatically compares the real-time driving trajectory data point-by-point with the location information, physical boundary range, and trigger judgment rules of the electronic fence; when the vehicle's positioning coordinates fall within the electronic fence area, it calculates the comprehensive risk value using the formula in Example 1. And determine whether it satisfies If the conditions are met, the entry is deemed valid. For vehicles deemed to have entered validly, their electronic fence entry time, continuous stay duration, and corresponding high-risk location information are marked, and the judgment result is sent to the alarm generation and push module 500. The judgment result and trajectory data of the trajectory acquisition and judgment module 400 are synchronously stored in the data storage module.

[0061] The alarm generation and push module 500, connected to the trajectory acquisition and judgment module 400, generates an alarm signal and pushes it to the law enforcement personnel's terminal when a monitored vehicle effectively enters the electronic fence area and its export consignment information indicates general cargo. Specifically, this module performs the following operations: based on the judgment result sent by the trajectory acquisition and judgment module 400, it automatically generates an alarm signal containing the vehicle's license plate number, vehicle monitoring time interval, electronic fence entry time, continuous stay duration, corresponding high-risk location name and geographical coordinates, container number, EIR information screenshot, and the vehicle's current real-time location; it pushes the alarm signal in real-time to the designated law enforcement personnel's mobile terminal and the monitoring backend terminal via a wired network or 5G / 4G wireless network; simultaneously, it automatically retains a complete alarm record in the monitoring backend, including the alarm generation time, alarm details, information on the involved vehicle and container, and the law enforcement personnel's reception and processing status. The alarm signal and alarm record are synchronously stored in the data storage module.

[0062] The enforcement feedback and dynamic update module 600 is connected to the high-risk location information database establishment module 100, the electronic fence deployment module 200, and the alarm generation and push module 500, respectively. It receives on-site inspection results from enforcement personnel and dynamically updates and optimizes the high-risk location information database and risk assessment model parameters based on these results. Specifically, this module performs the following operations: receiving on-site inspection results uploaded by enforcement personnel via terminal devices. Inspection results include confirmation of false declarations of dangerous goods transportation, no false declarations of dangerous goods found, or other abnormal transportation behaviors discovered. It also simultaneously records the complete trajectory information, stopping location, stopping time, and risk score results of the corresponding vehicles. When the inspection result confirms a false declaration, it automatically extracts the coordinates and trajectory characteristics of the actual loading location. If the location is not included in the high-risk location information database, it automatically adds a location record, achieving dynamic expansion of the information database. Based on the enforcement feedback results, it statistically analyzes the distribution characteristics of each risk value and dynamically adjusts the spatial risk weights. Time risk weight Speed ​​risk weight Historical behavior risk weights and risk alarm threshold The confidence adjustment parameters are dynamically adjusted based on the false alarm ratio and the missed alarm ratio. Increase when the false alarm rate rises. The value decreases when the proportion of missed alarms increases. The updated high-risk location information and optimized model parameters are synchronized to the high-risk location information database establishment module 100 and the electronic fence deployment module 200 for application in the next round of vehicle risk identification. The update records and optimized parameter sets of the enforcement feedback and dynamic update module 600 are synchronously stored in the data storage module.

[0063] The aforementioned modules work together to form a complete monitoring system. In a typical monitoring cycle, the high-risk location information database establishment module 100 and the electronic fence deployment module 200 first complete the configuration of the deployment and triggering rules for the monitored area; the monitoring list generation module 300 acquires multi-source data in real time and generates a list of monitored vehicles; the trajectory acquisition and judgment module 400 acquires vehicle trajectories based on the list and makes valid entry judgments; the alarm generation and push module 500 generates and pushes alarm signals for vehicles that have been judged to be valid; after law enforcement officers conduct on-site inspections, the law enforcement feedback and dynamic update module 600 receives the inspection results and performs dynamic updates to the information database and model parameters. The updated parameters are immediately applied to the next round of identification, realizing continuous self-optimization of the system.

[0064] Example 3 To verify the risk credibility correction factor constructed in this invention To assess the effectiveness of identifying false declarations in customs declarations for dangerous goods containers, this embodiment selects three typical regulatory scenarios and compares the risk assessment models without the risk credibility correction mechanism with those using the risk credibility correction mechanism of this invention.

[0065] The assessment model that does not use a risk credibility correction factor adopts the following comprehensive risk calculation method: Using risk credibility correction factor The risk assessment model adopts the following comprehensive risk calculation method: Among them, the risk credibility correction factor The above formula defines whether a vehicle exhibits a dangerous goods loading behavior pattern of "entering a dangerous goods area and continuously staying there while operating at low speed".

[0066] Scenario 1: The vehicle is passing through the hazardous materials factory area normally.

[0067] A container truck, while on a general cargo transport mission, was traveling normally along a road surrounding a hazardous materials factory. The vehicle was located in close proximity to the factory, posing a significant spatial risk. Relatively high; however, the vehicle always maintains a normal driving speed (e.g., v(t) is much greater than the reference speed). , (Low), no stay occurred (stay time T approaches zero, (Approaching zero).

[0068] In risk assessment models that do not employ risk credibility correction factors, a higher spatial risk value is achieved because only a linear weighted sum of each risk term is applied. Directly leads to the overall risk value A significant increase in risk level raises the possibility of it being misjudged as abnormal transportation behavior. However, in risk assessment models employing risk credibility correction factors, although... High, but time risk value and speed risk value All are at very low levels. According to the above formula, the risk credibility correction factor... Maintaining a low value close to zero effectively suppressed the final overall risk value. This prevented it from reaching the alarm threshold. Therefore, this invention can effectively prevent vehicles that normally pass through hazardous materials areas from being mistakenly identified as vehicles falsely reported as transporting hazardous materials.

[0069] Scenario 2: Vehicles temporarily stop due to traffic congestion or traffic signal control.

[0070] A container truck, while traveling on a road near a hazardous materials factory, experienced brief stops or slow, creeping movements due to traffic congestion or traffic light control. During these incidents, the vehicle's instantaneous speed decreased significantly, raising the speed risk value. Increased; simultaneously, the increased dwell time due to parking also increases the time risk value. It has also increased.

[0071] In risk assessment models that do not employ risk credibility correction factors, the combined risk value is higher due to the combined effect of speed risk and time risk. The risk level is significantly increased, which could easily lead to a false alarm. However, the vehicle did not actually enter the hazardous materials handling area, and its trajectory was some distance from the core operating area of ​​the hazardous materials plant; therefore, the spatial risk value is [not specified]. Maintain a low level. In risk assessment models employing risk credibility correction factors, due to spatial risk values... The risk credibility correction factor is insufficient to support the establishment of a dangerous goods loading behavior pattern. Maintaining a low level allows for effective correction of the linear weighting component, ultimately resulting in a comprehensive risk value. The readings are still below the alarm threshold. Therefore, this invention can effectively distinguish between normal road traffic behaviors such as traffic congestion and waiting at traffic lights and hazardous materials loading operations, significantly reducing the system's false alarm rate.

[0072] Scenario 3: Falsely declaring dangerous goods for loading.

[0073] A container truck declared its export cargo to customs as general cargo. After picking up the empty container, it actually entered the core operating area of ​​a hazardous materials factory to load hazardous materials. The vehicle remained in the high-risk hazardous materials area for an extended period, maintaining low-speed movement or remaining stationary during the loading process.

[0074] In this scenario: the vehicle remains located in the core area of ​​a high-risk zone, posing a spatial risk value. The level remains consistently high; the duration of stay (T) is relatively long, indicating a high time risk value. The risk level increases with increasing dwell time; the risk level increases when the vehicle is in a state of low-speed operation or stationary for a long time. Significant increase. Due to the simultaneous rise and high levels of the three indicators—space risk, time risk, and speed risk—based on the risk credibility adjustment factor... The numerator Increase, making The value rapidly approaches 1. Therefore, the comprehensive risk value calculated by the risk assessment model using the risk credibility correction mechanism is... With linear weighted part The values ​​remained largely consistent until they eventually exceeded the preset alarm threshold. The system automatically generates risk warning information.

[0075] Figure 3 This diagram illustrates the verification and comparison of risk credibility correction factors, specifically showing the comparison of judgment results between risk assessment models that do not employ risk credibility correction mechanisms and those that do, under three typical regulatory scenarios.

[0076] Compared with risk assessment models that do not employ risk credibility correction mechanisms, risk assessment models that employ risk credibility correction mechanisms can provide the same or even more explicit early warning outputs when faced with real dangerous goods loading activities. At the same time, the risk credibility correction mechanism significantly suppresses the risk of false alarms in non-loading scenarios, achieving a lower false alarm rate while maintaining a high ability to identify false alarms.

[0077] In summary, this invention, by constructing a multidimensional coupled risk assessment model and introducing a risk credibility correction factor, can accurately identify typical composite behavioral patterns in false declarations of dangerous goods container exports, effectively distinguish between normal driving behavior and illegal loading behavior, significantly improve the identification accuracy, reduce the false alarm rate, and achieve adaptive optimization of the model through an iterative loop mechanism, thus possessing outstanding technical effects and industrial application value.

[0078] Although the present invention has been described in detail with reference to the accompanying drawings and preferred embodiments, the invention is not limited thereto. Various equivalent modifications or substitutions can be made to the embodiments of the invention by those skilled in the art without departing from the spirit and essence of the invention. Such modifications or substitutions should all fall within the scope of the invention, or any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the invention should be covered within the protection scope of the invention. Therefore, the protection scope of the invention should be determined by the scope of the claims.

Claims

1. A method for identifying false declarations in customs declarations for the export of dangerous goods containers, characterized in that, Includes the following steps: Establish a database of high-risk locations for false declarations in customs declarations of dangerous goods containers exported; Electronic fences are set up based on high-risk geographical locations in the high-risk location information database to complete the control of the regulatory area, and multiple trigger rules are configured. The multiple trigger rules include risk judgment conditions in four aspects: space, time, speed and historical entry frequency. The multiple trigger rules adopt a comprehensive risk assessment model that includes a risk credibility correction factor. The risk credibility correction factor is used to jointly correct spatial risk, time risk and speed risk to suppress false high-risk scores caused by a single risk factor. By connecting with customs export information data, transportation information on container vehicles, and terminal yard equipment handover information, based on the obtained information such as container cargo type, container vehicle license plate information, and container vehicle pick-up and return times in and out of the yard, container vehicles with export shipping information as general cargo are screened out, and a list of container vehicles that need to be supervised is formed. The driving trajectory of the monitored vehicle during the period from the empty container leaving the site to the loaded container entering the site is obtained, and the comprehensive risk assessment model including the risk credibility correction factor is used to determine whether the monitored vehicle has effectively entered the electronic fence range. If the monitored vehicle effectively enters the electronic fence area and its export consignment information indicates it is ordinary cargo, an alarm signal will be generated and pushed to the law enforcement personnel's terminal, prompting them to intercept and inspect the suspected illegal container.

2. The method for identifying false declarations in customs declarations for the export of dangerous goods containers according to claim 1, characterized in that, The establishment of a high-risk location database for false declarations in the export customs declaration of dangerous goods containers includes: High-risk geographical locations are selected based on historical violation records and the safety risk levels of hazardous chemical production and storage enterprises. The historical violation records include violations of one or more times within the past 5 years. The safety risk levels of hazardous chemical production and storage enterprises include red, orange, and yellow levels. The red level corresponds to a score below 60, the orange level corresponds to a score between 60 and 75, and the yellow level corresponds to a score between 75 and 90. Collect the geographic coordinates of each of the high-risk geographical locations to determine the physical boundaries of the work area, entrances and exits and supporting areas covered by each location. A high-risk location information database is constructed based on the geographic coordinates and physical boundary range, and a dynamic update function is configured for the high-risk location information database.

3. The method for identifying false declarations in customs declarations for the export of dangerous goods containers according to claim 1, characterized in that, The configuration of multiple trigger rules includes: Define the coordinates for determining whether a vehicle can effectively enter the electronic fence; The comprehensive risk threshold determination rule for a vehicle's valid entry into the electronic fence is set as follows: When a vehicle meets the coordinate triggering condition, the comprehensive risk value of the vehicle within the electronic fence is calculated. The calculation formula is: ； in, , , , These are spatial risk weights, temporal risk weights, velocity risk weights, and historical behavior risk weights, satisfying... ,and ; This is a risk credibility correction factor. Space risk items Defined as: ； Where T represents the total time the vehicle spends within the electronic fence. and Let represent the distances between the vehicle and the centers of the first and second high-risk areas at time t, respectively. , For the scale parameters of the corresponding region; Time risk items Defined as: ； in, It is a time scale constant; Speed ​​risk item Defined as: ； Where v(t) is the instantaneous speed of the vehicle at time t. The reference velocity constant; Historical risk items Defined as: ； Where N represents the total number of times the vehicle has entered high-risk areas in its history; Risk credibility correction factor The mathematical form of this is the ratio of the product of the spatial risk term, the temporal risk term, and the velocity risk term to the sum of the product and the dynamic credibility adjustment parameter, expressed as: ； in, This is a dynamic credibility adjustment parameter, whose value is updated periodically based on historical law enforcement feedback results; when At that time, it is determined that the target vehicle has effectively entered the electronic fence range, among which This is a preset risk threshold.

4. The method for identifying false declarations in customs declarations for the export of dangerous goods containers according to claim 1, characterized in that, The screening process identifies container vehicles whose export shipping information indicates general cargo, resulting in a list of container vehicles requiring supervision, including: By connecting with three-party data—customs export information, transportation vehicle information, and terminal yard equipment handover information—we can obtain real-time information such as cargo consignment declaration type, container number, carrier vehicle license plate, container vehicle departure time for empty containers, and return time for loaded containers. Filter out container vehicles whose cargo consignment declaration type is clearly dangerous goods, and retain container vehicles whose cargo consignment declaration type is general cargo; Link the empty container pick-up and empty container return records of the same carrier vehicle and mark the monitoring time interval of the vehicle, which is from the time of empty container pick-up to the time of empty container return. Based on the screening results and the regulatory time frame, a structured list of container vehicles requiring regulation is generated.

5. The method for identifying false declarations in customs declarations for the export of dangerous goods containers according to claim 4, characterized in that, The determination of whether the monitored vehicle has effectively entered the electronic fence range includes: The system connects to the positioning system of the monitored vehicles. Based on the license plate information of the carrier vehicles in the list of container vehicles that need to be monitored, it obtains the real-time driving trajectory data of the corresponding vehicles within their monitoring time interval. The trajectory data includes the real-time positioning coordinates and driving timestamps of the vehicles. The real-time driving trajectory data is compared point by point with the location information, physical boundary range and multiple triggering rules of the electronic fence. Based on the comparison results, it is determined whether the monitored vehicle has effectively entered the electronic fence area. If it is determined to have effectively entered, the electronic fence entry time, continuous stay duration, and corresponding high-risk geographical location information of the vehicle are marked.

6. The method for identifying false declarations in customs declarations for the export of dangerous goods containers according to claim 4, characterized in that, The generation of the alarm signal and its push to the law enforcement officer's terminal includes: When it is determined that a monitored vehicle has effectively entered the electronic fence area and its export shipment declaration type is general cargo, an alarm signal for suspected false declaration is automatically generated; the alarm signal includes at least the vehicle's license plate, vehicle monitoring time interval, electronic fence entry time, continuous stay duration, corresponding high-risk geographical location information, screenshot of container equipment handover document information, and real-time vehicle location. The alarm signal is pushed to the terminal device of the designated law enforcement personnel via wired or wireless communication, and a complete alarm record is stored in the monitoring background. The alarm record includes the alarm generation time, alarm details, information on the vehicle and container involved, and processing status.

7. The method for identifying false declarations in customs declarations for the export of dangerous goods containers according to claim 3, characterized in that, Also includes: Receive on-site inspection results uploaded by law enforcement officers, including confirmation of false reporting, no false reporting or other abnormal transportation behavior found; When the inspection result confirms a false report, the coordinates and trajectory features of the loading location are extracted. If the location is not included in the high-risk location information database, the location is automatically added and the geographical coordinates and physical boundaries are collected, so as to realize the dynamic expansion of the information database. Based on the inspection results, we statistically analyze the valid alarm samples and false alarm samples, analyze the distribution characteristics of spatial risk values, temporal risk values, speed risk values, and historical behavior risk values, and dynamically adjust the weight of each risk item and the alarm threshold. The confidence adjustment parameter is dynamically adjusted based on the false alarm ratio and the missed alarm ratio. Increase when the false alarm rate rises. The value decreases as the percentage of missed alarms increases. value; The updated high-risk location information and optimized model parameters are synchronized to the regulatory system and applied to the next round of vehicle trajectory determination and risk identification, forming a self-feedback iterative mechanism.

8. The method for identifying false declarations in customs declarations for the export of dangerous goods containers according to claim 3, characterized in that, The risk credibility correction factor Joint characterization of spatial risk, temporal risk, and velocity risk: When spatial risk, temporal risk, and velocity risk occur simultaneously and all remain at a high level... The value approaches 1; When only a single abnormal risk characteristic exists The value remains at a low level.

9. The method for identifying false declarations in customs declarations for the export of dangerous goods containers according to claim 1, characterized in that, The high-risk geographical locations include high-risk hazardous materials factories, high-risk hazardous materials storage yards, and other high-risk locations for hazardous materials operations.

10. A system for identifying false declarations in customs declarations for the export of dangerous goods containers, characterized in that, include: The High-Risk Location Information Database Establishment Module is used to establish a high-risk location information database for false declarations of dangerous goods container export customs declarations. The electronic fence deployment module is used to set up electronic fences based on high-risk geographical locations in the high-risk location information database, complete the deployment of the monitored area, and configure multiple trigger rules. The multiple trigger rules include risk judgment conditions in four aspects: space, time, speed, and historical entry count. The multiple trigger rules adopt a comprehensive risk assessment model that includes a risk credibility correction factor. The risk credibility correction factor is used to jointly correct the spatial risk, time risk, and speed risk to suppress false high-risk scores caused by a single risk factor. The regulatory list generation module is used to connect with customs department's cargo export information data, transportation department's container vehicle information, and terminal yard equipment handover form information. Based on the container cargo type, container vehicle license plate information, and container vehicle pick-up and return time in and out of the yard in the obtained information, it screens out container vehicles whose export consignment information is general cargo and forms a list of container vehicles that need to be regulated. The trajectory acquisition and determination module is used to acquire the driving trajectory of the monitored vehicle from the time the empty container leaves the site to the time the loaded container enters the site, and to use the comprehensive risk assessment model containing the risk credibility correction factor to determine whether the monitored vehicle has effectively entered the electronic fence range. The alarm generation and push module is used to generate an alarm signal and push it to the law enforcement personnel's terminal when the monitored vehicle effectively enters the electronic fence range and its export consignment information is ordinary goods. The law enforcement feedback and dynamic update module is used to receive the on-site inspection results from law enforcement officers and to dynamically update and optimize the high-risk location information database and risk assessment model parameters based on the inspection results.

Images