Police work system based on real person authentication technology and implementation method

Abstract

The application provides a public security work system based on real person authentication technology and an implementation method, and belongs to the field of internet information technology.The system specifically comprises three function support modules: a background management research and judgment module, a real person authentication module and a data interface module;three business handling modules: a traffic violation handling module, a population flow management module and a real name clue reporting module;the background management research and judgment module serves as the core of the system, performs data interaction and scheduling with the three business handling modules through the data interface module, and is responsible for data statistical analysis and task distribution;the user real person authentication module provides identity verification services for the business handling module, and ensures that the identity information of the operator is reliable;the system and the implementation method can solve the problems existing in current public security work, such as traffic violation behavior deduction, untimely population information update and online reporting uncertainty.

Description

Technical Field

[0001] This invention belongs to the field of Internet information technology, specifically relating to a public security work system and its implementation method based on real-person authentication technology. Background Technology

[0002] Real-person authentication is an identity verification service that relies on biometric technologies such as liveness detection and facial recognition, combined with document OCR recognition technology, to accurately verify the identity of natural persons and enterprises, effectively ensuring the authenticity of the operator's identity.

[0003] LBS (Location Based Services) technology uses various positioning technologies to obtain the current location of a positioning device and provides information resources and basic services to the positioning device through the mobile Internet. Its core process is: users determine their own spatial location through positioning technology and then obtain location-related resources and information through the mobile Internet.

[0004] The public security work system is the core information technology carrier that supports daily law enforcement, government services and public order control. Its core functions are divided into business processing functions and data management functions.

[0005] The business processing functions include: public government services such as registration and management of the floating population, handling of traffic violations, and acceptance of tip-offs, as well as internal law enforcement business such as case investigation and personnel management; The data management function covers the collection, storage, query, and basic analysis of internal and external business data, providing data support for business processing.

[0006] In the actual operation of the existing public security system, there are obvious shortcomings in the implementation of the above-mentioned core functions. In particular, in key areas such as the management of the floating population that relies on location information and the handling of traffic violations that require accurate identity verification, the lack of technical solutions leads to low business processing efficiency and poor control accuracy.

[0007] The management of the floating population: The core premise of the management of the floating population is to accurately and in real time grasp the geographical location information and living dynamics of the people. The existing technology mainly relies on two methods to collect the geographical location information of the floating population: one is manual declaration and registration, that is, the floating population himself or his landlord or employer submits the living information to the local police station, or the police conduct door-to-door investigation and registration; the other is scattered collection based on a single positioning method, such as obtaining general area information only through the positioning of mobile phone operator base stations, or relying on users to actively upload location information.

[0008] The existing data collection methods have many insurmountable drawbacks: manual reporting and registration rely on the active cooperation of personnel, which leads to problems such as untimely reporting and high omission rates. In addition, manual data entry is inefficient, prone to data errors, and cannot achieve dynamic tracking; the positioning accuracy of a single base station is low, with large errors, making it impossible to accurately determine the specific residential address and distinguish between short-term stays and long-term residences; and user-uploaded location information suffers from strong subjectivity, difficulty in verifying data authenticity, and discontinuous updates.

[0009] In addition, the existing system lacks the ability to deeply analyze the collected location information, and is unable to mine the patterns of population movement and identify potential migrant populations from massive amounts of scattered data. As a result, the public security department is in a passive and lagging state in grasping the dynamics of population movement, making it difficult to achieve precise control.

[0010] Traffic violation processing: Current vehicle violation processing is divided into on-site and off-site enforcement. While payments can mostly be completed online, driver's license point deductions, due to the long-term cumulative point system, often require in-person processing at the traffic management office. This leads to frequent instances of point buying and selling, making effective control difficult and failing to ensure that the penalized individual matches the driver's license holder. The core reason for this problem is the lack of an accurate online identity verification mechanism in the existing system, failing to achieve a strong link between "person, certificate, and incident," making it difficult to accurately identify the responsible party for the violation during the offline processing.

[0011] The tip-off reporting process: The existing online reporting system only requires registered users to fill in the reporting information and personal identity information to submit the report. It lacks an effective identity verification mechanism, which cannot guarantee the authenticity of the informant's identity, leading to a proliferation of false reports and increasing the difficulty for public security organs to screen clues. At the same time, it is difficult to hold users who make false reports legally accountable, and informants cannot check the progress of their reports, lacking follow-up feedback, which affects the public's enthusiasm for reporting.

[0012] In addition, the existing system lacks the ability to analyze the correlation between reported information and geographical location and personnel identity, and cannot quickly locate the jurisdiction and key control points corresponding to the reported clues, resulting in low efficiency in the flow of clues and affecting the timeliness of case investigation.

[0013] In summary, the following specific problems still exist in the process of handling public security work, particularly regarding vehicle violation management, population flow information, and police-citizen information exchange: 1. The lack of an identity verification mechanism leads to insufficient business credibility: Existing systems, when handling traffic violations and tip-offs requiring clear identification of responsible parties, rely heavily on traditional identity verification methods such as usernames and passwords, and SMS verification. This makes it difficult to achieve robust verification of the consistency between the person, the document, and the machine. Particularly in the non-on-site penalty process for traffic violations, the lack of an effective real-person binding mechanism for online driver's license point deductions makes it difficult to prevent proxy point deductions, seriously undermining the fairness of law enforcement. In online reporting scenarios, the identity of the informant cannot be effectively verified, leading to frequent false and malicious reports. This not only wastes police resources but also affects the authenticity and traceability of leads.

[0014] 2. The management of the floating population relies on passive reporting, resulting in poor data timeliness and accuracy. Traditional methods of managing the floating population rely primarily on manual registration, reports from employers, or self-reporting by the floating population, resulting in significant issues of outdated information and missed registrations. Even when using single positioning methods such as cell tower triangulation, low positioning accuracy, inability to distinguish between long-term and short-term residences, and discontinuous data updates make it difficult to provide reliable data support for dynamic control. Consequently, public security departments struggle to grasp the true state of population movement in a timely manner, leading to passive management and delayed early warnings.

[0015] 3. Weak location information collection and analysis capabilities make it difficult to support accurate judgment: Existing systems often rely on single sources for location information collection, resulting in limited data accuracy and poor stability in complex environments such as indoor spaces and densely populated urban areas. The collected location data is often discrete and isolated, lacking effective cleaning, fusion, and clustering analysis methods. This makes it impossible to automatically identify people's habitual residences, workplaces, and movement patterns from massive amounts of coordinates, and even more difficult to conduct spatial correlation analysis with cases and key areas, thus hindering the implementation of intelligence-driven precision policing.

[0016] 4. The business systems suffer from severe data silos, resulting in low collaboration efficiency: Internal police systems such as traffic management, population management, and case reporting are often built independently, with inconsistent data standards and non-interoperable interfaces, creating "information silos." Business processes rely on manual transmission and secondary data entry, preventing data sharing and reuse. This lack of cross-domain data support hinders analysis and decision-making, impacting police response speed and the effectiveness of coordinated action.

[0017] 5. Inadequate data security and privacy protection mechanisms: During the data collection, transmission and use process, existing systems have relatively weak protection measures for personal sensitive information, with problems such as insufficient encryption strength, non-standard data desensitization and lax access control. These issues pose risks of data leakage and misuse, and make it difficult to meet increasingly stringent legal and regulatory compliance requirements. Summary of the Invention

[0018] In view of this, the present invention proposes a public security work system and implementation method based on real-person authentication technology, aiming to solve core problems in the process of public security work, such as deduction of points for traffic violations, untimely updates of information on the floating population, and inability to verify the authenticity of online reports.

[0019] To achieve the above objectives, the present invention adopts the following technical solution: This invention provides a public security work system based on real-person authentication technology, the system specifically including: a function support module and a business processing module; Specifically, the functional support module includes a back-end management and analysis module, a real-person authentication module, and a data interface module; The back-end management and analysis module serves as the core of the system, used for data statistical analysis, task distribution, and visualization. The real-person authentication module serves as the foundation of the system, used to verify user identity through biometric technology to ensure the authenticity of the operator's identity; The data interface module serves as the system's neural network, used for secure data exchange between various modules within the system and with external authoritative systems. Specifically, the business processing module includes a traffic violation processing module, a population flow management module, and a real-name clue reporting module; The traffic violation processing module is used to process traffic violation cases and binds the processing behavior with the real-person authentication personnel information; The population mobility management module is used to collect and analyze user geographic location information after authorization in order to manage the mobile population; The real-name tip-off module is used to receive tip-off information submitted by whistleblowers who have been verified by real-name authentication, and to bind the information to the whistleblower's identity; The back-end management and analysis module interacts and schedules data with each business processing module through the data interface module.

[0020] Furthermore, the data interface module includes internal interfaces and external interfaces; The external interface is used to interact with external systems. The external interface uses the SSL / TLS encryption protocol to build a secure transmission channel and performs desensitization processing on the transmitted personal sensitive information. The internal interface is used to coordinate the data flow between various modules within the system. The internal interface uses the AES-256 encryption algorithm to encrypt the data transmitted or stored internally, and sets data format verification rules and time stamp synchronization mechanisms.

[0021] Furthermore, the population flow management module includes: The positioning unit is used to obtain high-precision coordinate data of the user using the LBS multi-source positioning fusion method; The analysis engine includes a big data analysis engine and a rules engine. The big data analysis engine is used for data preprocessing, deep mining, and pattern recognition of location information. The rules engine is used for logical judgment, early warning triggering, and task generation based on preset rules.

[0022] This invention also provides a method for implementing a public security work system based on real-person authentication technology, the method specifically including the following steps: Step S1, Real Person Authentication: Users complete real-name authentication through methods such as facial recognition and liveness detection to ensure consistency between "person, certificate, and machine"; the system records user identity information as the identity basis for all subsequent business processes.

[0023] Real-person authentication provides a genuine identity for all subsequent business processes, eliminating fraudulent identity operations from the source and ensuring that all business activities are traceable to specific individuals.

[0024] Step S2, data acquisition and processing: With user authorization and in compliance with laws and regulations, the system collects user location coordinates in real time or at regular intervals using multi-source positioning technology, including but not limited to collecting user IP address, GPS positioning, operator base station positioning, etc., and processes the collected coordinates to improve positioning accuracy and stability. When collecting data, a reliable time period is preset, including time period one: 8 pm to 1 am; time period two: 2 am to 6 am.

[0025] By collecting data, a stable, continuous, and high-precision sequence of user location coordinates is obtained. Through data processing, high-quality data is obtained, providing reliable data for subsequent analysis.

[0026] Step S3, Spatial Pattern Perception and Structured Conversion: Receive the specific location information packets collected in Step S2, process the location information packets carrying multiple types of data, and transform the discrete, spatiotemporally disordered original coordinate sequences into a set of spatial clusters that are machine-recognizable and business-interpretable. Then, combined with the big data engine and rule engine, complete the geographic address conversion through spatial clustering analysis, high-risk area early warning, and spatial correlation analysis, convert the spatial cluster set into specific business features, and generate specific intelligence by associating it with other data, summarizing and outputting a personal mobility trend report.

[0027] Through transformation, disordered coordinate points are converted into spatial clusters with business significance, and personal mobility status reports are generated, achieving a qualitative change from data to intelligence.

[0028] Step S4, Business Status Determination and Task Generation: The system receives the personal mobility status report output in step S3, and accurately matches the report content with the rules according to the preset final business status classification rule library to finally determine the user's mobility management status and generate specific pending task instructions accordingly. The final business status classification rule library includes rules for determining whether a user is a newly added migrant population, rules for determining whether a user has left their original registration location, and rules for determining whether a user has changed their residential address.

[0029] The system automatically identifies statuses such as "newly added migrant population" and "change of residence," transforming manual screening into proactive system alerts and automatically generating structured tasks, greatly improving the efficiency and accuracy of analysis.

[0030] Step S5, Task Push and Visualization: The system automatically generates to-do tasks based on the analysis results of step S4 and associates them with user identity and location information; the tasks are visualized through a two-dimensional GIS map, marking the task location, type, and priority; and are pushed to the police officers of the corresponding police station through the data interface module, who can view the task status in detail through the two-dimensional GIS map.

[0031] Step S6, Task Feedback: The police officer receives the task pushed from Step S5, selects to verify by phone or in person according to the task priority, and updates the person's actual residence information in the system. The system updates the population database simultaneously, supporting subsequent queries and statistical analysis, and finally completes the closed loop.

[0032] The offline verification results are fed back to the system, forming a complete closed loop of "collection-analysis-push-verification-feedback-update". This ensures the real-time nature and accuracy of the population database and provides a reliable basis for subsequent statistical analysis.

[0033] Furthermore, the data acquisition and processing procedure in step S2 includes: Step S21, Multi-source data synchronous acquisition: Synchronously acquire raw positioning data packets with timestamps, coordinate values ​​and accuracy parameters from multiple positioning sources; Step S22, Environmental Adaptive Judgment and Dynamic Weight Allocation: Evaluate the current environmental characteristics, dynamically calculate and allocate fusion weights to each positioning source based on the preset environmental adaptation coefficient; Step S23, Adaptive weighted fusion and trajectory correction: Using a filtering algorithm, weighted multi-source data is fused to generate high-precision fused coordinates, and historical trajectory data is combined to smooth out abnormal jumps; Step S24, Data Integration Output: Output a high-precision location coordinate sequence for a continuous time period, including timestamps, latitude and longitude, positioning accuracy, and reliable time period labels.

[0034] Furthermore, the geographic location transformation analysis in step S3 specifically includes: Step S31, Data Preparation and Geographic Information Transformation: By utilizing the reverse geocoding function of the geographic information system, the latitude and longitude coordinates output in step S2 are reverse-looked into structured text addresses, and the machine-readable coordinate data is converted into human-readable and analyzable geographic tags, laying the foundation for subsequent analysis based on address text and spatial location. Step S32, Spatial Cluster Analysis: Cluster analysis is performed on the geographic tags output in step S31 using a big data engine and a rule engine: The geographic labels output in step S31 are transformed into clean, feature-rich, standardized data through data preprocessing. By configuring and optimizing parameters, the core parameters of the clustering algorithm are dynamically determined and optimized, so that the clustering results not only conform to the objective laws of data, but also meet the needs of public security management. Through algorithm execution and iteration, stable active clusters are identified from preprocessed data, and core points, boundary points and noise points are distinguished. Through result correction, verification, and output, the clustering results are modified in terms of business logic, evaluated in terms of quality, and output in a structured manner to ensure that the clustering analysis results can be directly used for the assessment of the floating population and to support the generation of subsequent tasks. Step S33, High-risk area warning: The clustering analysis results output in step S32 are subjected to spatial overlay analysis to associate the location with public security control elements. Once the system detects that the activity clusters of personnel have spatial intersection with high-risk areas, it automatically generates early warning information and pushes it to the local police or command center in real time. Step S34, Spatial correlation analysis: The spatial co-occurrence and movement patterns are transformed into analytical clues to serve core criminal investigation tasks such as case linkage and gang detection; the specific analysis content includes: Trajectory similarity analysis: By comparing the movement trajectories of different individuals, we can identify those who frequently travel together and whose routes highly overlap. Spatiotemporal co-occurrence analysis: Analyzing whether different people frequently appear in the same location at the same time; Social network analysis: Based on co-occurrence and trajectory similarity, a network graph of personnel associations is constructed.

[0035] Furthermore, the clustering algorithm execution process in step S32 includes: Step S321, Data preprocessing: The raw data output from step S2 is filtered and cleaned, specifically including: Outlier detection: The big data engine identifies abnormal coordinate jumps and performs outlier detection and calibration on the raw data output in step S2. Missing value imputation: missing values ​​are imputed at the breakpoints that occur throughout the monitoring process. For short-term chain breaks, linear interpolation or spline interpolation is used to fill in the missing values. Noise filtering: By applying a sliding window midpoint filtering algorithm through the big data engine, the slight coordinate jitter caused by signal reflection is smoothed out, thus completing the noise filtering of the original data; Define reliable time periods: Define the second period from 2:00 AM to 6:00 AM as the nighttime rest period; define the first period from 8:00 PM to 1:00 AM as the evening home period; define the period with continuously changing location and the same time within a detection cycle as the commuting period; define the business significance of fixed time periods to facilitate subsequent business analysis; Assign weights to different time periods: set the weight for nighttime rest periods to 1.2, the weight for evening home periods to 1.0, the weight for commuting periods to 0.8, and the weight for other time periods to 0.5. Apply these weights directly to feature weighting. Step S322, Parameter Configuration and Optimization: The core parameters are determined and dynamically adjusted using a rule engine: Neighborhood radius ε is determined based on city type: 30 meters for first-tier cities, 50 meters for second-tier cities, and 100 meters for county-level areas, ensuring that location points in the same residential / work area can be grouped into the same cluster; Minimum number of points (minPts) is determined by combining the statistical period setting. For a 30-day data period, minPts is set to 5; for a 60-day data period, minPts is set to 8; and for a 90-day data period, minPts is set to 12 to avoid short-term dwell points being misjudged as stable clusters. Density-based dynamic adjustment of ε: The rule engine receives local density values ​​provided by the big data engine. If the density of a certain region is more than twice the average, the ε of that region will be reduced to 80%; if the density is less than half the average, the ε of that region will be increased to 120%. Based on user type, the minPts is adjusted elastically: For those marked as key personnel or low-frequency activity personnel, if the effective location points within 30 days are less than 60% of the minPts, the minPts will be reduced to 3, and an extended period review mechanism will be enabled to avoid missing effective cluster information. Step S323, Algorithm Execution and Iteration: The big data engine is used to perform collaborative execution of the two algorithms. Based on the specific parameters determined in step S322, cluster analysis is performed on the preprocessed coordinate data. First, the K-means algorithm is used for preliminary clustering to quickly divide candidate clusters and reduce the computational difficulty of the DBSCAN algorithm. Then, the DBSCAN algorithm is executed separately for each candidate cluster to further split dense sub-clusters, remove noise points, and complete the fine clustering. The coordinate data is processed using a rule engine according to the following rules: Determine the core point: If the ε-neighborhood of a certain location point contains at least minPts points, and the average time-period weight of these points is ≥0.8, then it is determined to be the core point and is used as the core unit of the cluster; Cluster expansion: Using the core point ε, points located in the neighborhood of ε that satisfy the minPts condition are grouped into the same cluster to form a complete location cluster; Noise labeling: Points that cannot be classified into any cluster are labeled as noise points, and their location and time information are recorded for subsequent abnormal behavior analysis; Iterative optimization: Combining the clustering results of two consecutive statistical periods, if the overlap rate of the core points of a cluster is ≥70%, it is determined to be a stable cluster; if the overlap rate is <30%, the clustering process is re-executed to ensure the stability and accuracy of the cluster. Step S324, Result Correction, Verification, and Output: The rule engine formulates transformation and verification rules for spatial features to business features, checks the transformed results, and verifies the compliance of the checked results. Through the constraints of the transformation and verification rules, it ensures that the output cluster information does not contain precise geographic information that is prohibited from being disclosed by laws and regulations, and performs generalization processing when necessary. The big data analytics engine provides quantitative assessment and support services, organizing each cluster that has completed verification and correction, combining it with actual map markers, and visually displaying it through the data interface module.

[0036] Furthermore, the conversion verification rules mentioned in step S324 specifically include: Cluster merging rule: If the distance between the core points of two clusters is less than 2ε and the similarity of user behavior patterns within the clusters is greater than 80%, then the two clusters will be merged into one cluster; Cross-module verification rules: Population and household registration data and historical registration information from the backend management and analysis module are called through the data interface to verify the clustering results; if the residential cluster address identified by the clustering is consistent with the user's household registration address, the cluster is assigned a household registration label; if it overlaps with the area involved in the case, an association warning is triggered. The quantitative assessment and support services mentioned in step S324 specifically include: Cluster quality assessment: Calculate the silhouette coefficient and DBI index to evaluate intra-cluster compactness and inter-cluster separation, and generate a quantitative report on cluster quality; Business common sense verification: Verify whether the cluster distribution conforms to the common sense of public security business. If the residential cluster is distributed in a non-residential area or the work cluster is distributed in a non-commercial area / industrial park, it is determined to be an invalid cluster. Evaluation result correction: If the distance between the core points of two adjacent clusters is less than 2ε, and the difference between the user dwell time and the time distribution entropy within the cluster is less than 20%, then the two clusters will be merged; Noise point reclassification: If the location data marked as noise points appear more than or equal to 3 times within a reliable time period, they will be re-included in the clustering range and the neighborhood radius ε will be adjusted. If the residential cluster matches the user's registered address, it is marked as the registered address cluster; if it matches the historical registered address of the migrant population, the validity of the cluster is verified.

[0037] Furthermore, the data calculation and processing procedure in step S321 is as follows: Let P1 and P2 be two positioning points, and the specific attributes of P1 and P2 include: Coordinates: P1 is (lat1, lon1); P2 is (lat2, lon2); Positioning accuracy radius: P1 is R1, P2 is R2, the unit is meters; Timestamps: P1 is t1, P2 is t2, in seconds; Maximum permissible speed: Vmax, in meters per second.

[0038] Step S3211, calculate the spherical distance between P1 and P2: The basic distance Dbase between two points is calculated using the Haversine formula, in meters: ; ; ; Where Rearth is the Earth's radius, approximately 6,371,000 meters; Step S3212: Based on the base distance Dbase obtained in step S3211, calculate the distance range value after error adjustment. Minimum distance: At this point, it is assumed that the two error circles overlap the most as they face each other. Maximum distance: At this point, it is assumed that the two error circles are farthest apart in opposite directions; Step S3213: Based on the distance range obtained in S3212, calculate the possible velocity range: Time difference: The unit is seconds; Minimum speed: The unit is meters per second; Maximum speed: The unit is meters per second; Step S3214: Based on the speed range obtained in step S3213, determine whether the logic is abnormal according to the rules set by the rule engine. Main judgment rule: If If the process from P1 to P2 is deemed unrealizable, then P2 is considered an outlier. Here, k is the safety factor, with a value less than 1.2. The value of k varies depending on the region: 1.0–1.2 for suburban areas and highways, and 0.8–1.2 for urban roads. Auxiliary judgment rule: If R1 or R2 is greater than a specific threshold, and Vmaxcalc>Vmax, but Vmin≤Vmax, then mark P2 as a low-quality point for further analysis in subsequent steps.

[0039] Furthermore, the new rules for the floating population added in step S4 are as follows: Condition 1: Based on the time feature extraction in the cluster analysis in step S321, select the set of coordinates that appear for ≥3 days within 30 days at the same latitude and longitude coordinates. If such coordinates exist, it indicates that the user has stable activity traces in the area, and proceed to the subsequent address determination; if not, proceed to condition 5. Condition 2: Based on the weight allocation in step S321 and the algorithm judgment rules in step S323, the core point judgment in the clustering algorithm must meet the requirement that the average weight of the time period is ≥0.8. The clustering priority of time period 2 is the highest and the credibility is the strongest. Therefore, the number of days each address appears in time period 2 is counted first. If an address appears for ≥3 days and is the address with the most appearances in that time period, it is recorded as a credible address A, and the address is directly determined to be the user's permanent residence, terminating the subsequent conditions. If there is no address that meets the requirement of ≥3 days during the night rest period, proceed to condition 3. Condition 3: If condition 2 is not met, the location data of time period 1 and time period 2 are merged according to the spatial cluster merging rules in step S324 to ensure the consistency of the comprehensive judgment; the cumulative number of days each address appears in the two time periods is counted, and the address with the most days of appearance is recorded as the trusted address B, which is determined to be the user's permanent residence, and the subsequent conditions are terminated; if there is still no clear dominant address after merging, proceed to condition 4. Condition 4: When two or more trusted addresses are selected simultaneously by conditions 2 and 3, relying on the timestamp synchronization mechanism provided by the data interface, and combined with the cross-module verification rules of the clustering results provided in step S324, other addresses are excluded, and the location address of the most recently collected trusted time period is recorded as trusted address C, and the user's latest permanent residence is locked first. Condition 5: If condition 1 is not met, the statistical period is extended twice. The first extension is to 40 days. If condition 1 is still not met, the period is extended to 50 days. After each extension, the screening process of conditions 1 to 4 is repeated to ensure coverage of low-frequency location users. The location data within the extended period still needs to be cleaned and clustered. Step S322 is used to lower the minimum number of points threshold for low-frequency data to ensure that no effective clusters are missed. Condition 6: If the corresponding address is still not met after the extended period of 50 days, the noise point is re-judged, and the location coordinates of all reliable time periods within 30-50 days are analyzed. The address with the most occurrences is recorded as the suspected address D. If the location point corresponding to the suspected address D appears ≥3 times and is concentrated in the reliable time period, the suspected address D is re-included in the clustering range, and the ε value is adjusted to expand the neighborhood radius to ensure that low-frequency but stable residential addresses are not missed. The above six conditions are in a progressive relationship. Once the preceding conditions are met, the following conditions will no longer be executed. That is, after the trusted address A is determined, the subsequent conditions will no longer be executed. Based on the above six conditions, the trusted address obtained through screening is compared with the user's registered address. If the address is inconsistent with the registered address and the user has not registered migrant population information on the public security intranet, the user is determined to be a newly added migrant population. If the address is consistent with the registered address, the user is determined to reside at the place of registered residence and is not included in the newly added migrant population. The determination result is then pushed to the corresponding local police station. The specific rules for the departure of the floating population are as follows: The number of days the collected address is outside the province where the user's registered residential address is located within 60 days is greater than or equal to 3 days, and there are no cross-provincial changes, and no coordinates of the user's registered residential address appear. 60-day calculation rule: The first occurrence of an address outside the province is counted as day 1; The specific rules for changes in the floating population are as follows: For registered migrant workers, the collected geographic location information is analyzed, and conditions one, two, three, and four above are applied to derive reliable addresses A, B, and C. It is then determined whether the geographic coordinates match the user's registered residence on the public security intranet. If they are within the jurisdiction of the same police station, the residence is considered unchanged; if they are not within the jurisdiction of the same police station, the determined address is sent to the corresponding police station. Beneficial effects

[0040] Compared with existing technologies, the beneficial effects of the public security work system and implementation method based on real-person authentication technology provided by this invention include: I. Eliminating the Risk of Identity Misuse at the Identity Authentication and Business Security Level: Relying on the "person, certificate, and machine" three-in-one real-person authentication mechanism that combines facial recognition, liveness detection, and comparison with the public security database, a highly reliable identity foundation is provided for traffic violation processing, clue reporting, and other businesses, eliminating problems such as driver's license point deduction on behalf of others and false reports from the source, and ensuring the legality and fairness of business processing.

[0041] Achieving a traceable closed loop for business operations: All business activities are strongly linked to real identities, forming an anti-tampering closed loop of "authentication-processing-point deduction" in traffic violation processing, and establishing a "real-name submission-progress tracking-accountability" mechanism in clue reporting, which not only strengthens the responsibility for violations and reports, but also protects users' right to know and builds an environment of mutual trust between the police and the public.

[0042] II. Aspects of the effectiveness of migrant population management: From passive registration to proactive intelligent management: Integrating multi-source positioning technologies such as GPS, BeiDou, Wi-Fi, and base stations, combined with adaptive weighted fusion and trajectory correction algorithms, it breaks through the limitations of single positioning accuracy and achieves high-precision, continuous location acquisition in complex environments; through spatial clustering analysis and rule engine, it automatically identifies stable residential clusters and mobility patterns, transforming manual screening into proactive system discovery and accurate early warning, significantly improving the efficiency of identifying the floating population.

[0043] Improve the accuracy and timeliness of registration: Through mechanisms such as weighted allocation of multiple reliable time periods, multi-cycle data verification, and re-judgment of noise points, the system can accurately distinguish between long-term residents and short-term stays, effectively avoiding underreporting and misreporting; Establish a closed loop of "collection-analysis-push-verification-feedback-update" to ensure real-time synchronization of the population database and solve the problems of untimely and lagging data in traditional manual reporting.

[0044] Strengthen high-risk control and correlation analysis: Automatic matching and early warning of activity clusters and high-risk areas are achieved through spatial overlay analysis. Combined with trajectory similarity, spatiotemporal co-occurrence analysis and social network construction, accurate analytical clues are provided for case linkage and gang discovery, helping to improve the quality and efficiency of criminal investigation.

[0045] III. Data Processing and Analysis Capabilities: High-precision data acquisition and cleaning: The LBS multi-source positioning fusion method is adopted, and a stable and high-precision position coordinate sequence is generated through environmental adaptive weight allocation and filtering algorithms. Based on technologies such as velocity constraints, physical model calculation, and sliding window filtering, outlier detection, missing value filling and noise filtering are achieved to ensure data quality.

[0046] Intelligent data transformation and analysis: Latitude and longitude coordinates are transformed into structured addresses through reverse geocoding, and discrete coordinates are transformed into spatial clusters with business significance through K-means and DBSCAN dual-algorithm collaborative clustering; the accuracy and adaptability of analysis results are improved through dynamic parameter optimization and cross-module verification.

[0047] Data-driven decision-making upgrade: The back-end management and analysis module integrates all business data, and through visualization and in-depth analysis, it realizes an intuitive presentation of the distribution of the floating population, reporting hotspots, and task processing status, promoting the transformation of police decision-making from relying on experience to relying on specific data and intelligent systems.

[0048] IV. System Collaboration and Security Compliance: Breaking down business data barriers: Through an integrated design that combines three functional support modules with three business processing modules, the system enables the sharing of data on traffic violations, population flow, and tip-offs. The back-end management and analysis module coordinates and schedules data through the data interface module, enhancing cross-business collaborative capabilities and solving the problems of data isolation and business disconnect in traditional systems.

[0049] Comprehensive data security protection: The data interface module distinguishes between internal and external interfaces, constructing a multi-layered security protection system to prevent the risk of leakage and tampering during data transmission and storage, and complies with personal information protection and network security regulations.

[0050] Adapt to diverse business needs: Through dynamic parameter adjustment, low-frequency user adaptation, and multi-time period weight allocation, the system is adapted to various scenarios such as first-tier cities / county areas, high-frequency / low-frequency activity personnel, and indoor / outdoor environments, thereby improving the system's versatility and practicality.

[0051] V. Police efficiency and service experience: Reduce manual workload: The system automatically completes the identification of the floating population, task generation, and accurate push. Police officers can view the tasks visually through GIS maps without manual investigation and data entry. The system automatically extracts key information from reported clues and transfers them to the corresponding jurisdiction, reducing the cost of manual screening and diversion.

[0052] Enhancing the public service experience: Traffic violation processing can be conveniently handled online, clue reporting provides a progress tracking channel, and the registration of the floating population can be completed intelligently without active declaration. While improving management efficiency, this reduces the cost of handling affairs for the public and enhances service satisfaction.

[0053] VI. Technical Architecture and Module Collaboration: By deeply integrating technical architecture and business processes, previously fragmented and isolated public security business modules are integrated into an organic whole characterized by data interoperability, business linkage, intelligent collaboration, and unified scheduling. Through the organic integration of functional support modules and business processing modules, data interoperability and collaborative scheduling for traffic, population, and reporting services are achieved, thereby improving the overall work efficiency and coordination capabilities of the public security system. Attached Figure Description

[0054] Figure 1 This is a diagram showing the overall system structure modules and their connections. Figure 2 This is a flowchart of the method steps; Figure 3 This is a flowchart of the rules for the population mobility management module. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0056] As shown in the figure, this invention provides a public security work system and implementation method based on real-person authentication technology. The work system specifically includes three functional support modules and three business processing modules. The functional support modules include a back-end management and analysis module, a real-person authentication module, and a data interface module; the business processing modules include a traffic violation processing module, a population flow management module, and a real-name tip-off module.

[0057] The back-end management and analysis module, as the core of the system, is responsible for data statistics and task distribution. It also completes data interaction and data flow scheduling with the traffic violation processing module, population flow management module, and real-name clue reporting module through the data interface module, thus managing the system in a coordinated manner.

[0058] Specifically, this involves: task distribution, and the back-end management and analysis module receiving push information from the population flow management module and the real-name clue reporting module, and sending the push information to the relevant jurisdiction units at the relevant locations; Visual displays, primarily in the form of maps, visually show the distribution of the floating population, hotspots for tip-offs, and task processing status; Data statistical analysis involves in-depth mining and analysis of all business data. It integrates business data from three major modules—traffic violation processing, population flow management, and real-name tip-off—as well as external data related to these business data, conducts comprehensive analysis from multiple perspectives, and issues early warnings to responsible personnel based on the analysis results.

[0059] As the foundation of the system, the real-person authentication module provides highly secure user identity verification services for the three major business processing modules. It typically combines technologies such as facial recognition, liveness detection, and comparison with the Ministry of Public Security's ID card database to ensure that the operator's "person, ID card, and mobile phone" are all identical, providing reliable identity information for other modules to process tasks.

[0060] The data interface module acts as the system's "neural network," involving data exchange and integration with all internal and external systems. It is primarily responsible for information push and completion feedback for the three major business processing modules.

[0061] The data interface module specifically includes external interfaces and internal interfaces. The external interfaces are responsible for exchanging information with authoritative external systems such as the "Ministry of Public Security Traffic Safety Integrated Service Management Platform" and the "Population Information Management System" to conduct real-time queries and verifications of violation data, driver's license information, and household registration information. The internal interfaces are responsible for coordinating the data flow between various modules within the system, providing real-person authentication results to business modules, and pushing geographical location data to the analysis center.

[0062] The specific details regarding internal and external interfaces include: External Interfaces: To ensure data transmission security and compliance, external data transmission uses SSL / TLS encryption protocols to build a secure transmission channel, preventing data from being eavesdropped on, subjected to man-in-the-middle attacks, or tampered with when transmitted over the public network. Simultaneously, data involving sensitive personal information undergoes anonymization processing, adhering to the "minimum necessity principle," transmitting only the minimum necessary information fields for business operations, fundamentally reducing the risk of data leakage.

[0063] During data interaction, consistency checks are performed synchronously. The MD5 message digest algorithm is used to verify data integrity, ensuring that the received data is consistent with the sent data and avoiding tampering or loss during data transmission.

[0064] Internal interfaces: Internal data transmission uses the AES-256 encryption algorithm to encrypt internally stored or transmitted data with high strength, preventing data leakage due to internal network intrusion, abuse of privileges, or loss of physical media.

[0065] For critical data flowing across modules, data format validation rules are set to ensure that the structure, type, value range, and encoding of data conform to predefined specifications when data flows across modules / systems, preventing the entry and exit of non-standard information. Simultaneously, a time-stamp synchronization mechanism ensures data time sequence consistency between multiple modules, avoiding situations where the order in which data arrives at different modules is disordered in asynchronous, concurrent, or distributed environments. This ensures that each module processes data based on the correct logical time sequence, supporting the smooth operation of the overall business process.

[0066] The traffic violation processing module is mainly responsible for the targeted handling of violations, payment, and driver's license point deductions. It links the driver's license with the person handling the violation, thus technically eliminating the phenomenon of "buying and selling points" for point deductions.

[0067] If the user confirms that the violation is true, the system will automatically bind the driver's license verified by the user to the current processing behavior and complete the point deduction; if the user raises an objection, the system will transfer the objection information and evidence to the back-end management and analysis module through the data interface module for secondary confirmation or transfer it to the traffic management office for processing.

[0068] The population mobility management module is mainly responsible for analyzing the household registration information, authentication geographical location and permanent residence information of real-person authenticated users. With the user's authorization and in compliance with laws and regulations, it collects the user's geographical location information periodically or triggered by the system, then compares the information and updates the residence registration information in the system.

[0069] The population flow management module uses LBS multi-source positioning fusion method to obtain high-precision coordinate data. This method integrates multiple positioning methods such as GPS, Beidou, Wi-Fi, and base station positioning, and achieves complementary optimization of multi-source positioning data through adaptive weighted fusion algorithm.

[0070] When the user is in an open outdoor environment, GPS and Beidou positioning are the primary methods. When the user is indoors or in a shady environment, the system automatically switches to a collaborative working mode of Wi-Fi positioning and base station positioning, and uses historical positioning trajectories for auxiliary correction to ensure the stability and accuracy of coordinate data in complex environments.

[0071] The population flow management module embeds a big data analytics engine and a rules engine, which work together to complete the entire process of cleaning, mining, analyzing, and judging location information and analysis results. The big data analytics engine is specifically responsible for data preprocessing, deep mining, pattern recognition, and predictive analysis. The rules engine is specifically responsible for rule matching, logical judgment, early warning triggering, and task generation.

[0072] The real-name tip-off module is mainly responsible for binding the tip-off information to the tip-off's real identity information, while also providing a tip-off progress tracking function to ensure that the public security organs can accurately trace the tip-off and establish contact, while protecting the tip-off's right to know and improving public participation and trust.

[0073] First, users need to complete real-person authentication through the real-person authentication module. After authentication, users fill in the reporting information through the real-name clue reporting module, and the system binds the reporting information with the reporter's information.

[0074] The system extracts key information from the reported information through the back-end management and analysis module, such as geographical location, time, and type of case. Then, the system transfers the information to the corresponding case-handling unit through the data interface module based on the extracted key information.

[0075] After the case is completed, the investigating unit will selectively provide feedback on the information, but the system will track and display the processing status of the clue at any time, and will also inform the system that the processing has been completed even if there is no specific feedback.

[0076] This invention provides a method for implementing a public security work system based on real-person authentication technology, with a population flow management module as the core and this system as the carrier. The specific steps of the method are as follows: Step S1, Real Person Authentication: Users complete real-name authentication through methods such as facial recognition and liveness detection to ensure consistency between "person, certificate, and machine"; the system records user identity information as the identity basis for all subsequent business processes.

[0077] Step S2, data acquisition and processing: With user authorization and in compliance with laws and regulations, the system collects user location coordinates in real time or at regular intervals using multi-source positioning technology, including but not limited to collecting user IP address, GPS positioning, operator base station positioning, etc., and processes the collected coordinates to improve positioning accuracy and stability. When collecting data, a reliable time period is preset, including time period one: 8 pm to 1 am; time period two: 2 am to 6 am.

[0078] Step S3, Spatial Pattern Perception and Structured Conversion: Receive the specific location information packets collected in Step S2, process the location information packets carrying multiple types of data, and transform the discrete, spatiotemporally disordered original coordinate sequences into a set of spatial clusters that are machine-recognizable and business-interpretable. Then, combined with the big data engine and rule engine, complete the geographic address conversion through spatial clustering analysis, high-risk area early warning, and spatial correlation analysis, convert the spatial cluster set into specific business features, and generate specific intelligence by associating it with other data, summarizing and outputting a personal mobility trend report.

[0079] Step S4, Business Status Determination and Task Generation: The system receives the personal mobility status report output in step S3, and accurately matches the report content with the rules according to the preset final business status classification rule library to finally determine the user's mobility management status and generate specific pending task instructions accordingly. The final business status classification rule library includes rules for determining whether a user is a newly added migrant population, rules for determining whether a user has left their original registration location, and rules for determining whether a user has changed their residential address.

[0080] Step S5, Task Push and Visualization: The system automatically generates to-do tasks based on the analysis results of step S4 and associates them with user identity and location information; the tasks are visualized through a two-dimensional GIS map, marking the task location, type, and priority; and are pushed to the police officers of the corresponding police station through the data interface module, who can view the task status in detail through the two-dimensional GIS map.

[0081] Step S6, Task Feedback: The police officer receives the task pushed from Step S5, selects to verify by phone or in person according to the task priority, and updates the person's actual residence information in the system. The system updates the population database simultaneously, supporting subsequent queries and statistical analysis, and finally completes the closed loop.

[0082] The data acquisition and processing in step S2 specifically includes: Step S21: Multi-source data is collected synchronously, and raw location data is obtained from multiple complementary positioning sources at the same time to address the limitations of a single technology in different scenarios.

[0083] A set of raw positioning data packets with timestamps, coordinate values, accuracy radius, and signal quality parameters from different sources are output through multiple positioning sources.

[0084] Step S22: Environmental adaptive judgment and dynamic weight allocation. The current environmental characteristics are intelligently evaluated, and appropriate weights are assigned to each location source to achieve optimal fusion.

[0085] The system analyzes the data quality and consistency of each location source in real time, assesses the quality of the user's environment, and determines the status of the user's environment based on the assessment results.

[0086] Using algorithms such as the analytic hierarchy process (AHP), the final weight of each location source is dynamically calculated based on a preset environmental adaptation coefficient table and the user's current environmental state.

[0087] Step S23, adaptive weighted fusion and trajectory correction, fuses the weighted multi-source data into an optimal, smooth, high-precision coordinate.

[0088] By using the Kalman filter algorithm, the original positioning data packet obtained in step S1 is fused with the weights of multiple positioning sources in the user's environment obtained in step S22, and a more accurate fused coordinate is output.

[0089] When a scene change or abnormal data jump is detected, the system uses algorithms such as particle filtering, combined with rich historical trajectory data, to smooth and correct the current point, eliminate unreasonable positional changes, and generate a continuous trajectory that is more in line with the actual movement pattern.

[0090] Step S24: Data integration and output. The data processed by weighting in step S23 is arranged and output according to the timestamp rules. The output data is a high-precision location coordinate sequence of the user for 30-90 consecutive days. Each data entry contains core fields such as timestamp, latitude and longitude, positioning accuracy radius, and reliable time period label.

[0091] Based on the above steps, taking the user's location process from outdoors to indoors as an example, the process of the population flow module acquiring user coordinate data is explained: Real-time acquisition of multi-source data: The user is in an open outdoor environment, and the system simultaneously receives: GPS: (39.9042°N, 116.4074°E), accuracy radius: 8 meters, signal-to-noise ratio: 45dB; BeiDou: (39.9045°N, 116.4072°E), accuracy radius: 6 meters, signal-to-noise ratio: 48dB; Wi-Fi: 3 APs detected, with fuzzy location (39.9039°N, 116.4081°E), accuracy radius: 25 meters; Base stations: 2 base stations for triangulation (39.9048°N, 116.4069°E), accuracy radius: 50 meters.

[0092] Environmental adaptive judgment: Conduct targeted assessment and analysis of the user's environment: Visible satellite count: GPS 9 / BeiDou 12 → Excellent; Number of Wi-Fi access points: 3 → Medium; Base station signal strength: -75dBm / -82dBm → relatively weak; Based on multiple location sources, a location consistency check is performed to determine that the user's location is outdoors. Based on the determined location, a decision is made to enable the outdoor optimization mode.

[0093] Adaptive weighted fusion computation: Based on the environmental adaptive judgment results, the weights under multiple location sources are dynamically calculated using the Analytic Hierarchy Process (AHP):

[0094] Where the quality factor = f (signal-to-noise ratio, signal stability, historical accuracy); The results were fused using Kalman filtering: Fusion coordinates = ∑(weight i × coordinate i) + motion model correction =0.324×GPS+0.342×BeiDou+0.112×Wi-Fi+0.096×Base Station+Speed ​​Prediction Correction Term Final result: (39.90435°N, 116.4073°E), calculation accuracy radius: 5.2 meters.

[0095] The detection scenario has changed: When a user moves from outdoors to indoors, their location coordinates change, and the states of different location sources change accordingly: Satellite signal-to-noise ratio decreases: GPS 18dB / BeiDou 15dB; Increased number of Wi-Fi access points and improved RSSI; The inertial sensor displays the change in movement speed.

[0096] Based on the above, it was determined that the location attributes of the monitoring personnel had changed, and the weight allocation was readjusted using a lookup table method:

[0097] The reassigned weights are as follows: GPS weight: 0.3 → 0.1 (reduced weight) BeiDou weighting: 0.3 → 0.1 (reduced weighting) Wi-Fi weight: 0.2 → 0.5 (increased) Base station weight: 0.2 → 0.3 (slight increase) Based on the changed weights, the coordinates of the monitoring personnel after the change in position are recalculated to obtain the fused coordinate data of the monitoring personnel after the change in position, and the movement trajectory of the monitoring personnel is inferred based on the left side before and after the position change.

[0098] Final result: (39.90425°N, 116.40725°E), calculation accuracy radius: 12.2 meters.

[0099] The specific steps of the analysis in step S3 include: Step S31: Data Preparation and Geographic Information Transformation By utilizing the reverse geocoding function of Geographic Information Systems (GIS), latitude and longitude coordinates can be reverse-looked into structured text addresses, converting machine-readable coordinate data into human-readable and analyzable geographic tags.

[0100] Each location record is accompanied by standardized "time-coordinate-address" information, laying the foundation for subsequent analysis based on address text and spatial location.

[0101] Step S32 Spatial Cluster Analysis: Cluster analysis is performed on the geographic tags output in step S31 using a big data engine and a rule engine: The geographic labels output in step S31 are transformed into clean, feature-rich, standardized data through data preprocessing. By configuring and optimizing parameters, the core parameters of the clustering algorithm are dynamically determined and optimized, so that the clustering results not only conform to the objective laws of data, but also meet the needs of public security management. Through algorithm execution and iteration, stable active clusters are identified from preprocessed data, and core points, boundary points and noise points are distinguished. By correcting, verifying, and outputting the results, the clustering results are modified in terms of business logic, evaluated in terms of quality, and output in a structured manner to ensure that the clustering analysis results can be directly used for the assessment of the floating population and to support the generation of subsequent tasks.

[0102] Step S33 High-risk area warning: The clustering analysis results output in step S32 are subjected to spatial overlay analysis to associate the location with public security control elements. Once the system detects that the activity clusters of personnel have spatial intersection with high-risk areas, it automatically generates early warning information and pushes it to the local police or command center in real time.

[0103] Step S34 Spatial correlation analysis: The spatial co-occurrence and movement patterns are transformed into analytical clues to serve core criminal investigation tasks such as case linkage and gang detection; the specific analysis content includes: Trajectory similarity analysis: By comparing the movement trajectories of different individuals, we can identify those who frequently travel together and whose routes highly overlap. Spatiotemporal co-occurrence analysis: Analyzing whether different people frequently appear in the same location at the same time; Social network analysis: Based on co-occurrence and trajectory similarity, a network graph of personnel associations is constructed.

[0104] Step S32, spatial clustering analysis, involves a clustering algorithm. This algorithm is one of the core analytical tools of the big data analysis engine, primarily used in the population flow management module, and also indirectly serves the overall data analysis of the back-end management and judgment module. Its core function is to automatically classify massive, unordered personnel location data through machine intelligence, extracting meaningful spatial and behavioral patterns, thereby realizing the transformation from "data" to "intelligence" and supporting the system's intelligent judgment and proactive early warning functions. The clustering algorithm of this invention primarily uses the DBSCAN algorithm, supplemented by the K-means algorithm for preliminary analysis. The specific implementation process of the clustering algorithm includes: Step S321: Data Preprocessing

[0105] The process transforms raw, noisy location data into clean, feature-rich, standardized data, providing high-quality input for subsequent clustering. A big data engine leads the data cleaning and feature extraction process; a rules engine provides data cleaning and feature extraction rules to assist the big data engine in completing the data processing.

[0106] The raw data output from step S2 is filtered and cleaned using a three-step method: outlier detection, missing value imputation, and noise filtering. By using a big data engine to identify abnormal coordinate jumps, outlier detection and calibration are performed on the raw data output in step S2.

[0107] For any breaks that occur during the entire monitoring process, missing values ​​are interpolated. For short-term chain breaks, linear interpolation or spline interpolation is used to complete the chain. By applying a sliding window midpoint filtering algorithm through a big data engine, the slight coordinate jitter caused by signal reflection is smoothed out, thus completing the noise filtering of the original data.

[0108] Feature optimization and extraction are performed on the cleaned data: Calculate the overall point density and the spatial distribution entropy of the overall points; Calculate the time during which consecutive points lie within a certain radius, the number of times they occur daily / weekly, and the temporal regularity to extract the temporal characteristics of the data.

[0109] Set cleaning thresholds for the data processing of the big data engine and provide configurable rule parameters: data segments that are missing for more than 30 minutes will not be imputed and will be processed directly in segments; records whose coordinate jumps exceed the city diameter in a single day will be automatically discarded.

[0110] Define the trusted time periods mentioned in step S2 and specifically mark the business significance of the trusted time periods: define the second time period from 2:00 AM to 6:00 AM as the night rest period, define the first time period from 8:00 PM to 1:00 AM as the evening home period, and define the time periods where the location changes continuously and the same time occurs within a detection cycle as the commuting period.

[0111] The weights for the above time periods are assigned as follows: the weight for the nighttime rest period is set to 1.2, the weight for the evening home period is set to 1.0, the weight for the commuting period is set to 0.8, and the weight for other time periods is set to 0.5. The weights for each time period are then directly applied to the feature weighting.

[0112] The entire process was subject to compliance checks to ensure that the data cleaning process did not violate the minimum necessary principle for personal information protection, and high-frequency trajectory data that was not essential for business operations was downsampled.

[0113] Step S322 Parameter Configuration and Optimization:

[0114] The optimal parameters for the clustering algorithm are dynamically configured to adapt to different users, regions, and data quality scenarios. Based on data distribution characteristics and business scenarios, the core parameters of the clustering algorithm are dynamically determined and optimized to ensure that the clustering results conform to objective data laws and meet the needs of public security management.

[0115] The big data engine provides a data situation profile, determines the core parameters of the clustering algorithm, and initializes the core parameters; the rule engine executes parameter decisions and dynamically adjusts the parameters.

[0116] Utilizing big data engines to analyze data and create situational profiles: The overall density distribution, time coverage, and precision distribution of the preprocessed data are analyzed, and the average point spacing and the proportion of data in different time periods are calculated to provide a data basis for parameter initialization.

[0117] The core parameters are determined and dynamically adjusted using a rule engine: Neighborhood radius ε: dynamically adjusted according to city type, set at 30 meters for first-tier cities, 50 meters for second-tier cities, and 100 meters for county-level areas, to ensure that location points in the same residential / work area can be grouped into the same cluster; Minimum number of points (minPts): Based on the statistical period setting, minPts is set to 5 for a 30-day data period, 8 for a 60-day data period, and 12 for a 90-day data period to avoid short-term dwell points being misjudged as stable clusters.

[0118] Initial cluster centers: The K-means++ algorithm is used to optimize the selection of initial centers to avoid local optima caused by random initialization.

[0119] Density-based dynamic adjustment of ε: The rule engine receives local density values ​​provided by the big data engine. If the density of a certain area is more than twice the average value, the rule "reduce the ε of the area to 80%" is triggered; if the density is less than half the average value, the rule "expand the ε of the area to 120%" is triggered.

[0120] Based on user type, the minPts elastic adjustment is as follows: For personnel marked as "key personnel" or "low-frequency activity personnel", if the effective location points within 30 days are less than 60% of the minPts, the minPts will be reduced to 3, and an extended period review mechanism will be enabled to avoid missing effective cluster information.

[0121] Step S323: Algorithm execution and iteration:

[0122] By employing a dual-algorithm approach and iterative optimization, stable active clusters are identified from preprocessed data, and core points, boundary points, and noise points are distinguished. Multiple rounds of iteration and stability verification ensure the reliability of the clustering results, while labeling noise points provides clues for anomaly behavior analysis.

[0123] The big data engine is responsible for the collaborative execution of the two algorithms to perform cluster analysis on the preprocessed coordinate data: first, the K-means algorithm is used for preliminary clustering to quickly divide candidate clusters and reduce the computational difficulty of the DBSCAN algorithm; then, the DBSCAN algorithm is executed separately for each candidate cluster to further split dense sub-clusters, remove noise points, and complete refined clustering.

[0124] The coordinate data is processed using a rule engine according to the following rules: Determine the core point: If the ε-neighborhood of a certain location point contains at least minPts points, and the average time-period weight of these points is ≥0.8, then it is determined to be the core point and is used as the core unit of the cluster; Cluster expansion: Using the core point ε, points located in the neighborhood of ε that satisfy the minPts condition are grouped into the same cluster to form a complete location cluster; Noise labeling: Points that cannot be classified into any cluster are labeled as noise points, and their location and time information are recorded for subsequent abnormal behavior analysis; Iterative optimization: Combining the clustering results of two consecutive statistical periods, if the overlap rate of the core points of a cluster is ≥70%, it is determined to be a stable cluster; if the overlap rate is <30%, the clustering process is re-executed to ensure the stability and accuracy of the cluster.

[0125] The dual-algorithm parallel execution optimizes the computation method: For massive user location data, a distributed parallel computing architecture is adopted, and the data is sharded according to user ID, and the location data of different users are distributed to different computing nodes; each computing node independently executes the clustering algorithm to generate local clustering results; the local results are aggregated through the central node, eliminating duplicate clusters across nodes and forming a globally unified clustering result, thereby improving computational efficiency.

[0126] Step S324: Result Correction, Verification, and Output:

[0127] The clustering results are modified according to business logic, assessed for quality, and output in a structured manner. They are then cross-validated with external data such as household registration and historical registration to ensure that they can be directly used for the assessment of the floating population and to support the generation of subsequent tasks.

[0128] The rule engine defines the rules for transforming spatial features into business features, as well as the rules for validating the transformed results: Cluster merging rule: If the distance between the core points of two clusters is less than 2ε and the similarity of user behavior patterns within the cluster is >80%, they will be automatically merged into the same cluster.

[0129] Cross-module verification: The rule engine calls external data such as population and household registration data and historical registration information from the backend management and analysis module through the data interface to verify the clustering results; if the residential cluster address identified by the cluster is consistent with the household registration address of the person, the residential cluster is assigned a household registration label; if it overlaps with a certain area involved in the case, an association warning is triggered.

[0130] Results compliance review: Ensure that the output cluster information does not contain precise geographic information that is prohibited from being disclosed by laws and regulations, and perform generalization processing when necessary.

[0131] The big data analytics engine provides quantitative assessment and support services: Clustering quality assessment calculation: The silhouette coefficient, DBI index, and other evaluation indicators used to evaluate intra-cluster compactness and inter-cluster separation in the final clustering results are calculated to provide a quantitative report on clustering quality. At the same time, the results are verified by combining the common sense of public security business, such as residential clusters should be mainly distributed in residential areas and work clusters should be mainly distributed in commercial areas / industrial parks. If the location of a cluster does not match the scenario, it is judged as an invalid cluster and the clustering parameters are re-optimized.

[0132] Final corrections are made based on the clustering quality assessment results: Cluster merging correction: If the distance between the core points of two adjacent clusters is less than 2ε, and the difference in user dwell time and time distribution entropy within the cluster is less than 20%, then they are merged into the same cluster; Noise point reclassification: For location data marked as noise points, if the frequency of occurrence is ≥3 times and is concentrated in a reliable time period, the data will be re-included in the clustering range, and the ε value will be adjusted to expand the neighborhood radius to avoid valid data being misclassified. Cross-module data verification: Combining population and household registration data and historical registration information from the backend management and analysis module, if the residential cluster obtained by clustering is consistent with the user's household registration address, it is marked as the "household registration location cluster"; if it is consistent with the historical registration address of the floating population, the validity of the cluster is verified.

[0133] Output results: Each cluster that has been verified and corrected is organized and visualized through the data interface module, combined with the actual map labels.

[0134] The data calculation and processing steps are as follows: Let P1 and P2 be two positioning points, and the specific attributes of P1 and P2 include: Coordinates: P1 is (lat1, lon1); P2 is (lat2, lon2); Positioning accuracy radius: P1 is R1, P2 is R2, the unit is meters; Timestamps: P1 is t1, P2 is t2, in seconds; Maximum permissible speed: Vmax, in meters per second.

[0135] Step S3211, calculate the spherical distance between P1 and P2: The basic distance Dbase between two points is calculated using the Haversine formula, in meters: ; ; ; Where Rearth is the Earth's radius, approximately 6,371,000 meters.

[0136] Step S3212: Based on the base distance Dbase obtained in step S3211, calculate the distance range value after error adjustment. Minimum distance: At this point, it is assumed that the two error circles overlap the most as they face each other. Maximum distance: At this point, it is assumed that the two error circles are farthest apart in opposite directions.

[0137] Step S3213: Based on the distance range obtained in S3212, calculate the possible velocity range: Time difference: The unit is seconds; Minimum speed: The unit is meters per second; Maximum speed: The unit is meters per second.

[0138] Step S3214: Based on the speed range obtained in step S3213, determine whether the logic is abnormal according to the rules set by the rule engine. Main judgment rule: If If the process from P1 to P2 is deemed unrealizable, then P2 is considered an outlier. Here, k is the safety factor, with a value less than 1.2. The value of k varies depending on the region: 1.0–1.2 for suburban areas and highways, and 0.8–1.2 for urban roads.

[0139] Auxiliary judgment rule: If R1 or R2 is greater than a specific threshold, and Vmaxcalc>Vmax, but Vmin≤Vmax, then mark P2 as a low-quality point for further analysis in subsequent steps.

[0140] The aforementioned calculation process, by integrating spatiotemporal constraints and positioning errors, constructs an intelligent mechanism for identifying and cleaning abnormal positioning data. Its core lies in utilizing the physical constraints of movement speed and positioning accuracy to identify and eliminate unreliable location points caused by signal drift, equipment malfunction, or data acquisition errors. By employing precise digital processing calculations, positioning data processing is improved from simple filtering based on empirical thresholds to intelligent cleaning based on kinematics and error theory. Through environmental classification and safety factor adjustment, the anomaly detection strategy is adapted to specific scenarios. Simultaneously, the automated and interpretable cleaning mechanism avoids the risk of data tampering due to human intervention.

[0141] The new rules for the floating population added in step S4 are as follows: Condition 1: Based on the time feature extraction in the cluster analysis in step S321, select the set of coordinates that appear for ≥3 days within 30 days at the same latitude and longitude coordinates. If such coordinates exist, it indicates that the user has stable activity traces in the area, and proceed to the subsequent address determination; if not, proceed to condition 5. Condition 2: Based on the weight allocation in step S321 and the algorithm judgment rules in step S323, the core point judgment in the clustering algorithm must meet the requirement that the average weight of the time period is ≥0.8. The clustering priority of time period 2 is the highest and the credibility is the strongest. Therefore, the number of days each address appears in time period 2 is counted first. If an address appears for ≥3 days and is the address with the most appearances in that time period, it is recorded as a credible address A, and the address is directly determined to be the user's permanent residence, terminating the subsequent conditions. If there is no address that meets the requirement of ≥3 days during the night rest period, proceed to condition 3. Condition 3: If condition 2 is not met, the location data of time period 1 and time period 2 are merged according to the spatial cluster merging rules in step S324 to ensure the consistency of the comprehensive judgment; the cumulative number of days each address appears in the two time periods is counted, and the address with the most days of appearance is recorded as the trusted address B, which is determined to be the user's permanent residence, and the subsequent conditions are terminated; if there is still no clear dominant address after merging, proceed to condition 4. Condition 4: When two or more trusted addresses are selected simultaneously by conditions 2 and 3, relying on the timestamp synchronization mechanism provided by the data interface, and combined with the cross-module verification rules of the clustering results provided in step S324, other addresses are excluded, and the location address of the most recently collected trusted time period is recorded as trusted address C, and the user's latest permanent residence is locked first. Condition 5: If condition 1 is not met, the statistical period is extended twice. The first extension is to 40 days. If condition 1 is still not met, the period is extended to 50 days. After each extension, the screening process of conditions 1 to 4 is repeated to ensure coverage of low-frequency location users. The location data within the extended period still needs to be cleaned and clustered. Step S322 is used to lower the minimum number of points threshold for low-frequency data to ensure that no effective clusters are missed. Condition 6: If the corresponding address is still not met after the extended period of 50 days, the location coordinates of all reliable time periods within 30-50 days are analyzed through noise point re-judgment. The address with the most occurrences is recorded as suspected address D. If the location point corresponding to suspected address D appears ≥3 times and is concentrated in the reliable time period, suspected address D is re-included in the clustering range, and the ε value is adjusted to expand the neighborhood radius to ensure that low-frequency but stable residential addresses are not missed. The above six conditions are in a progressive relationship. Once the preceding conditions are met, the following conditions will no longer be executed. That is, after the trusted address A is determined, the subsequent conditions will no longer be executed. Based on the above six conditions, the trusted address obtained through screening is compared with the user's registered address. If the address is inconsistent with the registered address and the user has not registered migrant population information on the public security intranet, the user is determined to be a newly added migrant population. If the address is consistent with the registered address, the user is determined to reside in the place of registered residence and is not included in the newly added migrant population. The determination result is then pushed to the corresponding local police station.

[0142] The specific rules for the departure of the floating population are as follows: The number of days the collected address is outside the province where the user's registered residential address is located within 60 days is greater than or equal to 3 days, and there are no cross-provincial changes, and no coordinates of the user's registered residential address appear. The 60-day calculation rule is that the first occurrence of an address outside the province is counted as day 1.

[0143] The specific rules for changes in the floating population are as follows: For registered migrant workers, the collected geographic location information is analyzed, and conditions one, two, three, and four above are applied to derive reliable addresses A, B, and C. It is then determined whether the geographic coordinates match the user's registered residence on the public security intranet. If they are within the jurisdiction of the same police station, the residence is considered unchanged; if they are not within the jurisdiction of the same police station, the determined address is sent to the corresponding police station.

[0144] The specific content pushed out in step S5 includes: The user information and trusted address coordinates are pushed to the police station's backend system. After verification, the police can register the migrant population.

[0145] The police did not take any action. After 30 days of statistical analysis, this user was found to have registered on the public security intranet, and the information was pushed for deletion.

[0146] After the police handled the case, and after 30 days of statistical analysis, there was still no user registration information on the public security intranet, so the information was continued to be sent to the police station.

[0147] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A public security work system based on real-person authentication technology, characterized in that, include: Functional support module and business processing module; Specifically, the functional support module includes a back-end management and analysis module, a real-person authentication module, and a data interface module; The back-end management and analysis module serves as the core of the system, used for data statistical analysis, task distribution, and visualization. The real-person authentication module serves as the foundation of the system, used to verify user identity through biometric technology to ensure the authenticity of the operator's identity; The data interface module serves as the system's neural network, used for secure data exchange between various modules within the system and with external authoritative systems. Specifically, the business processing module includes a traffic violation processing module, a population flow management module, and a real-name clue reporting module; The traffic violation processing module is used to process traffic violation cases and binds the processing behavior with the real-person authentication personnel information; The population mobility management module is used to collect and analyze user geographic location information after authorization in order to manage the mobile population; The real-name tip-off module is used to receive tip-off information submitted by whistleblowers who have been verified by real-name authentication, and to bind the information to the whistleblower's identity; The back-end management and analysis module interacts and schedules data with each business processing module through the data interface module.

2. The public security work system based on real-person authentication technology according to claim 1, characterized in that, The data interface module includes internal interfaces and external interfaces; The external interface is used to interact with external systems. The external interface uses the SSL / TLS encryption protocol to build a secure transmission channel and performs desensitization processing on the transmitted personal sensitive information. The internal interface is used to coordinate the data flow between various modules within the system. The internal interface uses the AES-256 encryption algorithm to encrypt the data transmitted or stored internally, and sets data format verification rules and time stamp synchronization mechanisms.

3. The public security work system based on real-person authentication technology according to claim 1, characterized in that, The population mobility management module includes: The positioning unit is used to obtain high-precision coordinate data of the user using the LBS multi-source positioning fusion method; The analysis engine includes a big data analysis engine and a rules engine. The big data analysis engine is used for data preprocessing, deep mining, and pattern recognition of location information. The rules engine is used for logical judgment, early warning triggering, and task generation based on preset rules.

4. A method for implementing a public security work system based on real-person authentication technology, characterized in that: The implementation method is applied to the public security work system based on real-person authentication technology as described in any one of claims 1-3, and the implementation method specifically includes the following steps: Step S1, Real Person Authentication: Users complete real-name authentication through methods such as facial recognition and liveness detection to ensure consistency between "person, ID, and device"; the system records user identity information as the identity basis for all subsequent business processes; Step S2, data acquisition and processing: With user authorization and in compliance with laws and regulations, the system collects user location coordinates in real time or at regular intervals using multi-source positioning technology, including but not limited to collecting user IP address, GPS positioning, operator base station positioning, etc., and processes the collected coordinates to improve positioning accuracy and stability; a reliable time period is preset when collecting data, including time period one: 8 pm to 1 am; time period two: 2 am to 6 am; Step S3, Spatial Pattern Perception and Structured Conversion: Receive the specific location information packets collected in Step S2, process the location information packets carrying multiple types of data, and transform the discrete, spatiotemporally disordered original coordinate sequence into a set of spatial clusters that are machine-recognizable and business-interpretable; then, combined with the big data engine and rule engine, complete the geographic address conversion through spatial clustering analysis, high-risk area early warning, and spatial correlation analysis, convert the spatial cluster set into specific business features, and generate specific intelligence by associating it with other data, summarizing and outputting a personal mobility trend report; Step S4, Business Status Determination and Task Generation: The system receives the personal mobility status report output in step S3, and accurately matches the report content with the rules according to the preset final business status classification rule library to finally determine the user's mobility management status and generate specific pending task instructions accordingly. The final business status classification rule library includes rules for determining whether a user is a newly added migrant population, rules for determining whether a user has left their original registration location, and rules for determining whether a user has changed their residential address. Step S5, Task Push and Visualization: The system automatically generates to-do tasks based on the analysis results of step S4 and associates them with user identity and location information; the tasks are visualized on a two-dimensional GIS map, with the task location, type and priority marked; the tasks are pushed to the police officers of the corresponding police station through the data interface module, and the police officers can view the task status in detail on the two-dimensional GIS map. Step S6, Task Feedback: The police officer receives the task pushed from Step S5, selects to verify by phone or in person according to the task priority, and updates the person's actual residence information in the system. The system updates the population database simultaneously, supporting subsequent queries and statistical analysis, and finally completes the closed loop.

5. The implementation method of the public security work system based on real-person authentication technology according to claim 4, characterized in that, The data acquisition and processing procedure in step S2 includes: Step S21, Multi-source data synchronous acquisition: Synchronously acquire raw positioning data packets with timestamps, coordinate values ​​and accuracy parameters from multiple positioning sources; Step S22, Environmental Adaptive Judgment and Dynamic Weight Allocation: Evaluate the current environmental characteristics, dynamically calculate and allocate fusion weights to each positioning source based on the preset environmental adaptation coefficient; Step S23, Adaptive weighted fusion and trajectory correction: Using a filtering algorithm, weighted multi-source data is fused to generate high-precision fused coordinates, and historical trajectory data is combined to smooth out abnormal jumps; Step S24, Data Integration Output: Output a high-precision location coordinate sequence for a continuous time period, including timestamps, latitude and longitude, positioning accuracy, and reliable time period labels.

6. The implementation method of the public security work system based on real-person authentication technology according to claim 4, characterized in that, The geographic location conversion analysis in step S3 specifically includes: Step S31, Data Preparation and Geographic Information Transformation: By utilizing the reverse geocoding function of the geographic information system, the latitude and longitude coordinates output in step S2 are reverse-looked into structured text addresses, and the machine-readable coordinate data is converted into human-readable and analyzable geographic tags, laying the foundation for subsequent analysis based on address text and spatial location. Step S32, Spatial Cluster Analysis: Cluster analysis is performed on the geographic tags output in step S31 using a big data engine and a rule engine: The geographic labels output in step S31 are transformed into clean, feature-rich, standardized data through data preprocessing. By configuring and optimizing parameters, the core parameters of the clustering algorithm are dynamically determined and optimized, so that the clustering results not only conform to the objective laws of data, but also meet the needs of public security management. Through algorithm execution and iteration, stable active clusters are identified from preprocessed data, and core points, boundary points and noise points are distinguished. Through result correction, verification, and output, the clustering results are modified in terms of business logic, evaluated in terms of quality, and output in a structured manner to ensure that the clustering analysis results can be directly used for the assessment of the floating population and to support the generation of subsequent tasks. Step S33, High-risk area warning: The clustering analysis results output in step S32 are subjected to spatial overlay analysis to associate the location with public security control elements. Once the system detects that the activity clusters of personnel have spatial intersection with high-risk areas, it automatically generates early warning information and pushes it to the local police or command center in real time. Step S34, Spatial correlation analysis: The spatial co-occurrence and movement patterns are transformed into analytical clues to serve core criminal investigation tasks such as case linkage and gang detection; the specific analysis content includes: Trajectory similarity analysis: By comparing the movement trajectories of different individuals, we can identify those who frequently travel together and whose routes highly overlap. Spatiotemporal co-occurrence analysis: Analyzing whether different people frequently appear in the same location at the same time; Social network analysis: Based on co-occurrence and trajectory similarity, a network graph of personnel associations is constructed.

7. The implementation method of the public security work system based on real-person authentication technology according to claim 6, characterized in that, The clustering algorithm execution process in step S32 includes: Step S321, Data preprocessing: The raw data output from step S2 is filtered and cleaned, specifically including: Outlier detection: The big data engine identifies abnormal coordinate jumps and performs outlier detection and calibration on the raw data output in step S2. Missing value imputation: missing values ​​are imputed at the breakpoints that occur throughout the monitoring process. For short-term chain breaks, linear interpolation or spline interpolation is used to fill in the missing values. Noise filtering: By applying a sliding window midpoint filtering algorithm through the big data engine, the slight coordinate jitter caused by signal reflection is smoothed out, thus completing the noise filtering of the original data; Define reliable time periods: define 2:00 AM to 6:00 AM as the nighttime rest period; define 8:00 PM to 1:00 AM as the evening home period; define time periods where the location changes continuously and the same time occurs within a detection cycle as the commuting period; define the business significance of fixed time periods to facilitate subsequent business analysis; Assign weights to different time periods: set the weight for nighttime rest periods to 1.2, the weight for evening home periods to 1.0, the weight for commuting periods to 0.8, and the weight for other time periods to 0.

5. Apply these weights directly to feature weighting. Step S322, Parameter Configuration and Optimization: The core parameters are determined and dynamically adjusted using a rule engine: Neighborhood radius ε is determined based on city type: 30 meters for first-tier cities, 50 meters for second-tier cities, and 100 meters for county-level areas, ensuring that location points in the same residential / work area can be grouped into the same cluster; Minimum number of points (minPts) is determined by combining the statistical period setting. For a 30-day data period, minPts is set to 5; for a 60-day data period, minPts is set to 8; and for a 90-day data period, minPts is set to 12 to avoid short-term dwell points being misjudged as stable clusters. Density-based dynamic adjustment of ε: The rule engine receives local density values ​​provided by the big data engine. If the density of a certain region is more than twice the average, the ε of that region will be reduced to 80%; if the density is less than half the average, the ε of that region will be increased to 120%. Based on user type, the minPts is adjusted elastically: For those marked as key personnel or low-frequency activity personnel, if the effective location points within 30 days are less than 60% of the minPts, the minPts will be reduced to 3, and an extended period review mechanism will be enabled to avoid missing effective cluster information. Step S323, Algorithm Execution and Iteration: The big data engine is used to perform collaborative execution of the two algorithms. Based on the specific parameters determined in step S322, cluster analysis is performed on the preprocessed coordinate data. First, the K-means algorithm is used for preliminary clustering to quickly divide candidate clusters and reduce the computational difficulty of the DBSCAN algorithm. Then, the DBSCAN algorithm is executed separately for each candidate cluster to further split dense sub-clusters, remove noise points, and complete the fine clustering. The coordinate data is processed using a rule engine according to the following rules: Determine the core point: If the ε-neighborhood of a certain location point contains at least minPts points, and the average time-period weight of these points is ≥0.8, then it is determined to be the core point and is used as the core unit of the cluster; Cluster expansion: Using the core point ε, points located in the neighborhood of ε that satisfy the minPts condition are grouped into the same cluster to form a complete location cluster; Noise labeling: Points that cannot be classified into any cluster are labeled as noise points, and their location and time information are recorded for subsequent abnormal behavior analysis; Iterative optimization: Combining the clustering results of two consecutive statistical periods, if the overlap rate of the core points of a cluster is ≥70%, it is determined to be a stable cluster; if the overlap rate is <30%, the clustering process is re-executed to ensure the stability and accuracy of the cluster. Step S324, Result Correction, Verification, and Output: The rule engine formulates transformation and verification rules to convert spatial features into business features, check the transformed results, and verify the compliance of the checked results. Through the limitation of the transformation and verification rules, it ensures that the output cluster information does not contain precise geographic information that is prohibited from being disclosed by laws and regulations, and performs generalization processing when necessary. The big data analytics engine provides quantitative assessment and support services, organizing each cluster that has completed verification and correction, combining it with actual map markers, and visually displaying it through the data interface module.

8. The implementation method of the public security work system based on real-person authentication technology according to claim 7, characterized in that, The conversion verification rules mentioned in step S324 specifically include: Cluster merging rule: If the distance between the core points of two clusters is less than 2ε and the similarity of user behavior patterns within the clusters is greater than 80%, then the two clusters will be merged into one cluster; Cross-module verification rules: Population and household registration data and historical registration information from the backend management and analysis module are called through the data interface to verify the clustering results; if the residential cluster address identified by the clustering is consistent with the user's household registration address, the cluster is assigned a household registration label; if it overlaps with the area involved in the case, an association warning is triggered. The quantitative assessment and support services mentioned in step S324 specifically include: Cluster quality assessment: Calculate the silhouette coefficient and DBI index to evaluate intra-cluster compactness and inter-cluster separation, and generate a quantitative report on cluster quality; Business common sense verification: Verify whether the cluster distribution conforms to the common sense of public security business. If the residential cluster is distributed in a non-residential area or the work cluster is distributed in a non-commercial area / industrial park, it is determined to be an invalid cluster. Evaluation result correction: If the distance between the core points of two adjacent clusters is less than 2ε, and the difference between the user dwell time and the time distribution entropy within the cluster is less than 20%, then the two clusters will be merged; Noise point reclassification: If the location data marked as noise points appear more than or equal to 3 times within a reliable time period, they will be re-included in the clustering range and the neighborhood radius ε will be adjusted. If the residential cluster matches the user's registered address, it is marked as the registered address cluster; if it matches the historical registered address of the migrant population, the validity of the cluster is verified.

9. The implementation method of the public security work system based on real-person authentication technology according to claim 4, characterized in that, The data calculation and processing procedure in step S321 is as follows: Let P1 and P2 be two positioning points, and the specific attributes of P1 and P2 include: Coordinates: P1 is (lat1, lon1); P2 is (lat2, lon2); Positioning accuracy radius: P1 is R1, P2 is R2, the unit is meters; Timestamps: P1 is t1, P2 is t2, in seconds; Maximum permissible speed: Vmax, in meters per second; Step S3211, calculate the spherical distance between P1 and P2: The basic distance Dbase between two points is calculated using the Haversine formula, in meters: ； ； ； Where Rearth is the Earth's radius, approximately 6,371,000 meters; Step S3212: Based on the base distance Dbase obtained in step S3211, calculate the distance range value after error adjustment. Minimum distance: At this point, it is assumed that the two error circles overlap the most as they face each other. Maximum distance: At this point, it is assumed that the two error circles are farthest apart in opposite directions; Step S3213: Based on the distance range obtained in S3212, calculate the possible velocity range: Time difference: The unit is seconds; Minimum speed: The unit is meters per second; Maximum speed: The unit is meters per second; Step S3214: Based on the speed range obtained in step S3213, determine whether the logic is abnormal according to the rules set by the rule engine. Main judgment rule: If If the process from P1 to P2 is deemed unrealizable, then P2 is deemed an outlier. Here, k is the safety factor, which is less than 1.

2. The value of k varies in different regions, ranging from 1.0 to 1.2 for suburban areas and highways, and from 0.8 to 1.2 for urban roads. Auxiliary judgment rule: If R1 or R2 is greater than a specific threshold, and Vmaxcalc>Vmax, but Vmin≤Vmax, then mark P2 as a low-quality point for further analysis in subsequent steps.

10. The implementation method of the public security work system based on real-person authentication technology according to claim 8, characterized in that, The specific rules for adding migrant population in step S4 are as follows: Condition 1: Based on the time feature extraction in the cluster analysis in step S321, select the set of coordinates that appear for ≥3 days within 30 days at the same latitude and longitude coordinates. If such coordinates exist, it indicates that the user has stable activity traces in the area, and proceed to the subsequent address determination; if not, proceed to condition 5. Condition 2: According to the weight allocation in step S321 and the algorithm judgment rules in step S323, the core point judgment in the clustering algorithm must meet the condition that the average weight of the time period is ≥0.

8. The clustering priority of time period 2 is the highest and the credibility is the strongest. Therefore, the number of days each address appears in time period 2 is counted first. If an address appears for ≥3 days and is the address that appears for the most days during that period, it is recorded as a trusted address A, and the address is directly determined to be the user's permanent residence, terminating the subsequent conditions; if there is no address that meets the condition of ≥3 days during the night rest period, proceed to condition three; Condition 3: If condition 2 is not met, the location data of time period 1 and time period 2 are merged according to the spatial cluster merging rules in step S324 to ensure the consistency of the comprehensive judgment; the cumulative number of days each address appears in the two time periods is counted, and the address with the most days of appearance is recorded as the trusted address B, which is determined to be the user's permanent residence, and the subsequent conditions are terminated; if there is still no clear dominant address after merging, proceed to condition 4. Condition 4: When two or more trusted addresses are selected simultaneously by conditions 2 and 3, relying on the timestamp synchronization mechanism provided by the data interface, and combined with the cross-module verification rules of the clustering results provided in step S324, other addresses are excluded, and the location address of the most recently collected trusted time period is recorded as trusted address C, and the user's latest permanent residence is locked first. Condition 5: If condition 1 is not met, the statistical period is extended twice. The first extension is to 40 days. If condition 1 is still not met, the period is extended to 50 days. After each extension, the screening process of conditions 1 to 4 is repeated to ensure coverage of low-frequency location users. The location data within the extended period still needs to be cleaned and clustered. Step S322 is used to lower the minimum number of points threshold for low-frequency data to ensure that no effective clusters are missed. Condition 6: If the corresponding address is still not met after the extended period of 50 days, the noise point is re-judged, and the location coordinates of all reliable time periods within 30-50 days are analyzed. The address with the most occurrences is recorded as the suspected address D. If the location point corresponding to the suspected address D appears ≥3 times and is concentrated in the reliable time period, the suspected address D is re-included in the clustering range, and the ε value is adjusted to expand the neighborhood radius to ensure that low-frequency but stable residential addresses are not missed. The above six conditions are in a progressive relationship. Once the preceding conditions are met, the following conditions will no longer be executed. That is, after the trusted address A is determined, the subsequent conditions will no longer be executed. Based on the above six conditions, the trusted address obtained through screening is compared with the user's registered address. If the address is inconsistent with the registered address and the user has not registered migrant population information on the public security intranet, the user is determined to be a newly added migrant population. If the address is consistent with the registered address, the user is determined to reside at the place of registered residence and is not included in the newly added migrant population. The determination result is then pushed to the corresponding local police station. The specific rules for the departure of the floating population are as follows: The number of days the collected address is outside the province where the user's registered residential address is located within 60 days is greater than or equal to 3 days, and there are no cross-provincial changes, and no coordinates of the user's registered residential address appear. 60-day calculation rule: The first occurrence of an address outside the province is counted as day 1; The specific rules for changes in the floating population are as follows: For registered migrant populations, the collected geographic location information is analyzed, and conditions one, two, three, and four above are executed to obtain reliable addresses A, B, and C. It is then determined whether the geographic coordinates match the user's registered residence on the public security intranet. If they are within the jurisdiction of the same police station, it is assumed that the residence has not changed. If the addresses are not under the jurisdiction of the same police station, the analysis will be sent to the corresponding police station.

Images