Access control event-driven dynamic management methods, devices, equipment, and media for personnel profiles

By generating event tuples from access control events and facial images, and combining them with gate topology and similarity calculations, the problem of inaccurate file creation for unregistered personnel and frequency statistics for the same person in access control management is solved. This enables real-name merging and anonymous file management, improving the accuracy and efficiency of management.

CN123115451APending Publication Date: 2026-11-17SHENZHEN PUHUI ZHIWEI TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611620100.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-10-20
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

In existing access control management systems, the separation of access control and image capture makes it impossible to create files for unregistered personnel. Pure video-based file aggregation lacks physical and spatiotemporal constraints. The frequency of the same person being split into multiple files is underestimated, or the frequency of different people being mistakenly combined and overestimated. The lack of address tags and verification work orders creates a closed loop, making it impossible to achieve one-to-one real-name management.

Method used

By acquiring access control event tuples and face image generation event tuples from the access controller, quality gate control detection is performed. The comprehensive similarity is calculated by combining the gate topology and face feature similarity. Dual-database matching and merging confidence determination are then performed to achieve real-name merging and dynamic management of stranger profiles.

Benefits of technology

It has achieved full-coverage management with one file per person, improved the accuracy of file merging, generated complete travel trajectories and verification work orders, and enhanced registration and archiving efficiency and abnormal behavior detection capabilities.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123115451A_ABST
    Figure CN123115451A_ABST
Patent Text Reader

Abstract

This invention discloses a method, apparatus, device, and medium for dynamic management of personnel files driven by access control events. The method acquires access control events and facial images to obtain event tuples. These event tuples are compared with both a resident database and a stranger database. Through a mechanism that combines similarity, dual-database matching, and merging confidence, the method separates the merging paths for real-name users from those for strangers. Finally, the resident files and stranger files, after real-name merging and trajectory aggregation, are integrated to form a unified personnel file. This invention preserves the identity certainty of real-name users, avoiding the merging of different individuals into one person, while also establishing traceable anonymous files for strangers through clustering technology, ultimately achieving the goal of one file per person and full-coverage personnel management.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of security monitoring technology, and in particular to a method, device, equipment and medium for dynamic management of personnel files driven by access control events. Background Technology

[0002] Smart communities, rental housing, and densely populated areas with high migrant populations commonly deploy access control systems at unit doors, community gates, and garage doors, simultaneously capturing facial images at these entrances. Management requires adherence to standardized addresses, actual population, actual housing units, and actual businesses: registration of arrivals, tracking of movement, and accurate data collection. However, in reality, two separate sets of data exist: the access control system records card swipes or facial recognition transactions, while the video system stores a massive number of captured images. Unregistered migrants are only detected after opening a door, but the system neither immediately creates a file nor combines multiple captures into a single person; community grid workers still rely on door-to-door registration. Inconsistent image quality, the splitting of the same person into multiple stranger files, and misattribution of different individuals distort trajectory and frequency statistics, preventing effective control and early warning systems from reaching the individual level.

[0003] The existing technologies for access control management mainly have the following problems: 1. Access control and image capture are separated, and no file is created when the door is opened: the passage flow and the image capture are stored in separate databases. Unregistered personnel will only have one failed verification or image capture, and no sustainable personal file will be created.

[0004] 2. Pure video aggregation lacks hard constraints in physical time and space. Relying solely on feature similarity, it is easy to mistakenly merge two people from the same field at the same second, or to mistake a snapshot that appears at an unreachable threshold within a short period of time as the same person; low-quality images participating in the voting will pollute the core of the archive.

[0005] 3. The establishment of files and the merging of records are not based on the same set of identity granularity. Merging of the passerby database, deduplication of employee files, and cross-regional identity archiving either assume that the identity is known or only organize the stranger database. It is impossible to complete the matching of the two databases, the creation of new files, and subsequent real-name promotion at the same time as opening the door.

[0006] 4. Trajectories and frequencies are attached to fragmented files. If the same person is split into multiple files, the frequency will be underestimated and warnings will be missed; if different people are mistakenly combined, the frequency will be overestimated and false alarms will be issued.

[0007] 5. The address tag and verification work order loop are missing. The standard address cannot be automatically transferred from the door to the person's file. Work orders that can be dispatched cannot be generated for abnormal visits or suspected unregistered residences.

[0008] Therefore, there is an urgent need for a dynamic personnel file management method driven by access control events. Summary of the Invention

[0009] In view of the above problems, the present invention is proposed to provide a method, apparatus, device and medium for dynamic management of personnel files driven by access control events to overcome or at least partially solve the above problems.

[0010] Other features and advantages of the invention will become apparent from the following detailed description, or may be learned in part by practice of the invention.

[0011] According to a first aspect of the present invention, a method for dynamic management of personnel files driven by access control events is provided, comprising: S1. Obtain access control events reported by the access control controller and capture facial images through the door point camera, and generate event tuples based on the access control events and facial images; S2. Perform quality gating detection on the event tuple. If the quality gating detection is passed, calculate the gate travel time based on the gate topology and calculate the comprehensive similarity by combining the facial feature similarity. S3. Based on the comprehensive similarity, perform dual-database matching on the resident database and the stranger database in descending order of priority to obtain the matching results; S4. Perform a file creation or update operation based on the matching result. If no match is found with either the resident file in the resident database or the stranger file in the stranger database, a new stranger file is created based on the current event tuple. If a match is found with either the resident database or the stranger database, the file core of the corresponding file is updated based on quality weighting. S5. When a stranger's profile matches a resident's profile in the resident database, calculate the merge confidence level and execute automatic merging or generate a merge approval work order for approval based on the preset merge threshold strategy. S6. Based on all merged resident files and stranger files, generate a unified personnel file, update the address tags corresponding to the personnel file and generate a travel trajectory, and generate a verification work order based on the travel frequency in the statistics window.

[0012] In some embodiments of the present invention, step S1, generating the event tuple includes: Based on the access control events reported by the access controller, obtain the door number d, access time t, credential identifier c, and event type τ; Obtain the face image output by the gate camera and extract the face feature vector x from it; Based on the face image, the corresponding sharpness s, pose p, illumination l, and relative width z of the face frame are obtained respectively. The quality score q of the face image is calculated by weighted summation of the sharpness s, pose p, illumination l, and relative width z of the face frame. Map the gate number d to the standard address a; The event tuple e is generated based on the gate number d, passage time t, credential identifier c, facial feature vector x, quality score q, event type τ, and standard address a.

[0013] In some embodiments of the present invention, step S2, which involves calculating the gate travel time based on the gate topology and combining it with facial feature similarity to calculate the overall similarity, includes: The gate topology is an undirected graph of gates. The shortest path walking time between two gates is calculated based on this undirected graph as the theoretical travel time. Determine whether the time difference between the current access control event and the historical access control event is less than the theoretical travel time minus the tolerance value; If so, the travel time is deemed unreachable, and the overall similarity score is set to invalid. If not, then the spatiotemporal affinity is calculated, which is determined based on the hierarchical relationship of the standard addresses corresponding to the current access control event and the historical access control events in the address tree; The comprehensive similarity is obtained by weighted summation of the spatiotemporal affinity and facial feature similarity.

[0014] In some embodiments of the present invention, step S3, which involves performing dual-database matching on the resident database and the stranger database in descending order of priority based on the comprehensive similarity, includes: Confirm whether the credential identifier in the event tuple matches the credential field in the resident database. If the match is successful, bind the current event tuple to the corresponding matching resident file. If the credential identifier fails to match the credential field in the resident database, it is confirmed whether the overall similarity is greater than the resident similarity threshold corresponding to the resident file in the resident database. If it is greater than the resident similarity threshold, the current event tuple is merged with the corresponding resident file for real-name authentication. If the overall similarity is not greater than the resident similarity threshold, then it is confirmed whether the overall similarity is greater than the stranger similarity threshold corresponding to the stranger profile in the stranger database. If the overall similarity is greater than the stranger similarity threshold, then the current event tuple is merged into the corresponding stranger profile.

[0015] In some embodiments of the present invention, step S5, where a real-name comparison between a stranger's profile and a resident profile in the resident database matches, calculates the merging confidence level and performs automatic merging or generates a merging approval work order for approval based on a preset merging threshold strategy, includes: The merge confidence level is obtained by weighted summation of the average similarity between the stranger profile and the resident profile, the co-occurrence ratio on common gate points or adjacent addresses, and the real-name hit identifier. If the merge confidence level is greater than or equal to the automatic merge threshold, then the stranger profile and the resident profile are merged, retaining the primary key of the resident profile; If the confidence level of merging is less than the automatic merging threshold but greater than or equal to the approval threshold, a merging approval work order is generated, and the stranger's file and the resident's file are merged after the merging approval work order is approved.

[0016] In some embodiments of the present invention, step S6, which involves updating the address tag corresponding to the personnel file and generating the travel trajectory, includes: For each of the aforementioned personnel files F, calculate the calculations at each standard address. a The event score (F, a) of the access control event is: ; In the formula, d e Let q be the gate point. e For quality score, t now -t represents the difference in passage time between two consecutive access control events, and γ represents the daily attenuation coefficient; Select the standard address 'a' with the highest event score and mark it as the main address of the personnel file F; Sort all event tuples in the personnel file F in ascending order of passage time to form a passage event sequence; From the sorted access event sequence, extract the door number of each access control event in sequence to obtain the door access sequence of the person; Using an undirected graph of gate points, find the shortest path between two adjacent gate points, and then concatenate the shortest paths between each pair of adjacent gate points to form a complete travel trajectory.

[0017] In some embodiments of the present invention, step S6, which involves generating a verification work order based on the frequency of passage within the statistical window, includes: If the personnel file is a stranger file and the number of times they arrive at any standard address within the long-term statistical window is greater than or equal to the high-frequency threshold, then an unregistered inflow verification work order will be generated. If the personnel file appears for the first time within a short-term statistical window and is matched with multiple passage times within a preset nighttime period, an unregistered residence verification work order will be generated; If the personnel file matches multiple non-adjacent standard addresses within a short-term statistical window, an abnormal wandering verification work order is generated.

[0018] According to a second aspect of the present invention, an access control event-driven dynamic personnel file management device is provided, the device comprising: The event binding module is used to obtain access control events reported by the access control controller, and to collect face images through the door point camera, and generate event tuples based on the access control events and face images; The constraint calculation module is used to perform quality gating detection on the event tuple. If the quality gating detection is passed, the gate travel time is calculated based on the gate topology, and the comprehensive similarity is calculated by combining the facial feature similarity. The dual-database matching module is used to perform dual-database matching on the resident database and the stranger database in descending order of priority based on the comprehensive similarity, and obtain the matching result; The file update module is used to perform file creation or update operations based on the matching results. If no match is found with either the resident file in the resident database or the stranger file in the stranger database, a new stranger file is created based on the current event tuple. If a match is found with either the resident database or the stranger database, the file core of the corresponding file is updated based on quality weighting. The merge decision module is used to calculate the merge confidence level when a stranger's profile matches a resident's profile in the resident database, and to perform automatic merging or generate a merge approval work order for approval based on a preset merge threshold strategy. The work order generation module is used to generate a unified personnel file based on all merged resident files and stranger files, update the address tags corresponding to the personnel file and generate the passage trajectory, and generate a verification work order based on the passage frequency in the statistics window.

[0019] According to a third aspect of the present invention, a computer device is provided, including a processor and a memory, the memory storing computer program instructions executable by the processor, wherein when the processor executes the computer program instructions, it implements the instructions as described in any of the above methods.

[0020] According to a fourth aspect of the present invention, a computer-readable storage medium is provided, wherein computer program instructions are stored therein, the computer program instructions being loaded and executed by a processor to perform the operations performed by the method described in any of the preceding claims.

[0021] The technical solutions provided in the embodiments of the present invention have at least the following technical effects or advantages: This invention provides a method, apparatus, device, and medium for dynamic management of personnel files driven by access control events. The method acquires access control events and facial images to obtain event tuples. These event tuples are compared with a resident database and a stranger database. Through a mechanism that combines similarity, dual-database matching, and merging confidence, the method separates real-name merging paths from stranger merging paths. Finally, the resident files and stranger files, after real-name merging and trajectory aggregation, are integrated to form a unified personnel file. This preserves the identity certainty of real-name personnel, avoids merging different people into one person, and establishes traceable anonymous files for strangers through clustering technology. Ultimately, it achieves the goal of one file per person and full-coverage personnel management.

[0022] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 A flowchart illustrating a method for dynamic management of personnel files driven by access control events, provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the principle structure of a personnel file dynamic management device driven by access control event provided in an embodiment of the present invention; Figure 3 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0025] Exemplary embodiments of this application will now be described in more detail with reference to the accompanying drawings.

[0026] The accompanying drawings illustrate various structural schematics according to embodiments of this application. These drawings are not to scale, and some details have been enlarged for clarity, and some details may have been omitted. The shapes of the various regions and layers shown in the drawings, as well as their relative sizes and positional relationships, are merely exemplary and may deviate from reality due to manufacturing tolerances or technical limitations. Furthermore, those skilled in the art can design regions / layers with different shapes, sizes, and relative positions as needed.

[0027] It should be noted that the terms "comprising" and "having," and any variations thereof, as used in this application, are intended to cover non-exclusive inclusion. The term "multiple" as used in this application refers to two or more. In the context of this application, similar or identical components may be denoted by the same or similar reference numerals.

[0028] To better understand the above technical solutions, the following will describe the above technical solutions in detail with reference to specific implementation methods. It should be understood that the embodiments of this application and the specific features in the embodiments are detailed descriptions of the technical solutions of the present invention, rather than limitations on the technical solutions of the present invention. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0029] Figure 1 This is a flowchart illustrating a method for dynamic management of personnel files driven by access control events, as provided in an embodiment of the present invention. Figure 1 As shown, the method for dynamic management of personnel files driven by access control events includes the following steps: S1. Obtain access control events reported by the access control controller and capture facial images through the door point camera, and generate event tuples based on the access control events and facial images; The event tuple generated in this embodiment of the invention includes: obtaining the door number d, access time t, credential identifier c, and event type τ based on the access control access event reported by the access control controller; the door number d is the door / device number reported by the access control controller, which corresponds to the primary key of the door address table; the access time t is the capture timestamp of the access control access event; the credential identifier c is the credential number (such as card number, account number, or personnel ID) returned by the access control controller when the user successfully opens the door by swiping a card, entering a password, or facial recognition; for unregistered strangers, the value of the credential identifier c is empty; the event type τ is defined according to the result of opening the door (whether it is successful access, verification failure, or being identified as a stranger), that is, the access control access event includes successful access event, verification failure event, and stranger event; Acquire the face image output by the gate camera and extract the face feature vector x from it; the face feature vector x is obtained by extracting features from the face image using a deployed (pre-trained) face feature model; Based on the face image, the corresponding sharpness s, pose p, illumination l, and relative width z of the face bounding box are obtained respectively. The quality score q of the face image is calculated by weighted summation of the sharpness s, pose p, illumination l, and relative width z. In the formula, w s w p w l wz Each of these represents a weight for a sub-item, for example, four weights of 0.35, 0.25, 0.25, and 0.15, summing to 1; the sharpness s is the sharpness of the face region, which can be scored using Laplacian variance or an existing sharpness model, and then normalized to [0,1] based on the historical high scores of that point; the pose p is the degree of yaw / pitch deviation from the frontal face estimated from facial key points (eyes, nose, etc.), for example, a frontal face is close to 1, and a large profile is close to 0; the illumination l is whether the brightness of the face region falls within the usable band, for example, too dark or overexposed is close to 0; the relative width z of the face bounding box is the ratio of the face bounding box width to the image width; The door number d is mapped to a standard address a; based on the preset door address table (which binds each door number to its corresponding standard address), the standard address a is obtained after obtaining the door number d. The standard address a is the physical address of the person's residence (e.g., a community, a building, a unit, or a unit number). The event tuple e is generated based on the gate number d, the passage time t, the credential identifier c, the facial feature vector x, the quality score q, the event type τ, and the standard address a, where e = (d, t, c, x, q, τ).

[0030] In other embodiments of the present invention, if the access control controller and the door point camera are deployed separately, then during the time window [ t δ , t + δ The best quality snapshot (highest quality score) from the camera at that gate location is compared with the event tuple. e Binding, δ Examples are 1 to 3 seconds.

[0031] S2. Perform quality gating detection on the event tuple. If the quality gating detection is passed, calculate the gate travel time based on the gate topology and calculate the comprehensive similarity by combining the facial feature similarity. This invention compares whether the quality score q in the event tuple is greater than or equal to the file update threshold q. min If the mass fraction q ≥ q min If the quality gate detection is successful, the image will pass the quality gate detection; otherwise, the quality gate detection will fail. Consequently, the resident database or stranger database will not be updated to avoid polluting the database with unclear, low-quality facial images. This invention's embodiments calculate gate travel time based on gate topology and combine it with facial feature similarity to calculate overall similarity, including: The gate topology is an undirected graph. The shortest path walking time between two gates is calculated based on this graph as the theoretical travel time. In the undirected graph, gates are treated as nodes, and walkable pathways are treated as edges. For example, the main gate of the residential complex is d0, the entrance to Unit 1 of Building 1 is d1, and the entrance to Unit 2 of Building 1 is d2. The edges are d0-d1 and d0-d2, respectively. If the two units are connected via the basement, d1-d2 is added. Nodes are derived from the gate address table, and edges are derived from the adjacency relationships of the residential complex plane or address tree. The walking time of each edge of the shortest path from any two gates di to dj in the undirected graph is extracted. When fellow students arrive =0, the edge weight can be the passage length divided by the walking speed (1.2m / s is commonly used in implementation), or written into the door point address table according to the empirical value of the address tree (e.g., 15s for the same unit, 45s for different units in the same building). Determine whether the time difference between the current access control event and the historical access control event is less than the theoretical travel time minus the tolerance value; If so, the travel time is deemed unreachable, and the overall similarity score is set to invalid; if... If the travel time is deemed unreachable, it is prohibited to classify them as belonging to the same person. `ti` and `tj` represent the travel times in the event tuples corresponding to the two gate points `di` and `dj`, respectively. slack This is the tolerance value for clock error and individual differences in step speed (e.g., 15s). If not, then the spatiotemporal affinity is calculated, which is determined based on the hierarchical relationship of the standard addresses corresponding to the current access control event and historical access control events in the address tree; if the two access points are the same access point in the address tree hierarchy, the spatiotemporal affinity is... ψ Take a first preset value (e.g., 1.0); if they are different gate points in the same unit, the spatiotemporal affinity is... ψ Take a second preset value (e.g., 0.7); if they are different units in the same building, the spatiotemporal affinity is... ψ Take a third preset value (e.g., 0.4); if they are different buildings in the same community, the spatiotemporal affinity is... ψ Take the fourth preset value (e.g., 0.2); if they are not in the same cell, then the spatiotemporal affinity is... ψ =0; A comprehensive similarity score is obtained by weighted summation of the spatiotemporal affinity and facial feature similarity; the comprehensive similarity score... α + β = 1, where α and β are the spatiotemporal affinity, respectively. ψ st and facial feature similarity s f The corresponding weights; the facial feature similarity s f Let x be the cosine similarity of the face feature vectors x in the two event tuples.

[0032] S3. Based on the comprehensive similarity, perform dual-database matching on the resident database and the stranger database in descending order of priority to obtain the matching results; The process of matching the resident database and the stranger database based on the comprehensive similarity, in descending order of priority, includes: The system confirms whether the credential identifier in the event tuple matches the credential field in the resident database. If the match is successful, the current event tuple is bound to the corresponding resident file. The resident file in the resident database contains multiple fields, such as the person's primary key (ID), credential field, and facial feature vector. The credential field is the credential identifier (card_id), such as the IC card / ID card number, password account, or the person's primary key returned after successful facial verification (i.e., the person's unique ID in the system). By precisely matching the credential identifier in the event tuple with the credential field in the resident database, if a resident file contains a credential field with the same credential identifier, the match is considered successful. Binding the current event tuple to the corresponding resident file means that this access control event is included in the historical log of the resident's corresponding resident file. Every card swipe / facial recognition access by this resident will be fully recorded to form a historical access trajectory. If the credential identifier fails to match the credential field in the resident database, it is confirmed whether the overall similarity is greater than the resident similarity threshold corresponding to the resident file in the resident database. If it is greater than the resident similarity threshold, the current event tuple is merged with the corresponding resident file for real-name authentication. If the overall similarity is not greater than the resident similarity threshold, then it is confirmed whether the overall similarity is greater than the stranger similarity threshold corresponding to the stranger profile in the stranger database. If the overall similarity is greater than the stranger similarity threshold, then the current event tuple is merged into the corresponding stranger profile, and the profile is updated with the current face feature vector x.

[0033] S4. Perform a file creation or update operation based on the matching result. If no match is found with either the resident file in the resident database or the stranger file in the stranger database, a new stranger file is created based on the current event tuple. If a match is found with either the resident database or the stranger database, the file core of the corresponding file is updated based on quality weighting. When no match is found between the resident profiles in the resident database and the stranger profiles in the stranger database, this embodiment of the invention creates a new stranger profile based on the current event tuple, with the current face feature vector x as the initial profile center, the cumulative quality weight = quality score q, the address label initially being the current standard address, and the passage trajectory corresponding to the current access control communication event added to the stranger profile; After a successful match with the resident database or the stranger database, this embodiment of the invention updates the corresponding file core based on quality weighting. The new profile core is obtained by weighting the existing profile core xˉ of the stranger profile with the current facial feature vector x according to their respective quality weights. .

[0034] S5. When a stranger's profile matches a resident's profile in the resident database, calculate the merge confidence level and execute automatic merging or generate a merge approval work order for approval based on the preset merge threshold strategy. In this embodiment of the invention, when a real-name comparison between a stranger's profile and a resident profile in the resident database matches, a merging confidence level is calculated, and automatic merging or merging approval work order is executed according to a preset merging threshold strategy, including: The merge confidence level is obtained by weighted summation of the average similarity between the stranger profile and the resident profile, the co-occurrence ratio on common gate points or adjacent addresses, and the real-name hit identifier. The combined confidence level In the formula, ρ represents the average similarity between the stranger profile and the resident profile. co I represents the co-occurrence ratio at common gate points or adjacent addresses. ld For real-name matching, λ1, λ2, and λ3 are the weights of the corresponding sub-items, and the sum of the weights is 1. For example, λ1 is 0.55, λ2 is 0.20, and λ3 is 0.25; the average similarity between the stranger profile and the resident profile. That is, after the quality scores of the stranger profile and the resident profile are both detected by quality gating, the comprehensive similarity of the two profiles is calculated and the average of the two is taken; the co-occurrence ratio ρ at the common gate or adjacent addresses. co The number of times events occur together at the same gate or adjacent addresses in the address tree in the stranger's file and the resident's file in their respective access trajectory tables (composed of a series of trajectory points, each trajectory point containing at least the access time, gate number, standard address, and event type t), is the proportion of the number of events in the two files that occur together, [0,1]; the real-name hit identifier I ld This involves checking whether the registered individual's ID image or registration information matches the profile of strangers. If a match is found, the real-name matching identifier I is used. ld It is 1 if it is true, otherwise it is 0.

[0035] If the merge confidence level is greater than or equal to the automatic merge threshold (e.g., 0.78), the stranger profile and the resident profile are merged, retaining the primary key of the resident profile; if the merge confidence level is less than the automatic merge threshold but greater than or equal to the approval threshold (e.g., 0.60), a merge approval work order is generated, and the stranger profile and the resident profile are merged after the merge approval work order is approved.

[0036] If a hard constraint conflict exists regarding travel time, automatic merging is prohibited. For example, if the calculated theoretical shortest travel time between two gates (e.g., 2 minutes) is less than the physical limit time required for humans to complete this journey in reality (e.g., 5 minutes), meaning the time interval between the current access control event and the historical access control event is less than the theoretical shortest travel time, then a hard constraint conflict regarding travel time is determined, and merging the current access control event into the corresponding file is prohibited.

[0037] S6. Based on all merged resident files and stranger files, generate a unified personnel file, update the address tags corresponding to the personnel file and generate a travel trajectory, and generate a verification work order based on the travel frequency in the statistics window.

[0038] This invention's embodiments of updating the address tags corresponding to the personnel files and generating travel trajectories include: for each personnel file F, calculating the address tags for each standard address. a The event score (F, a) of the access control event is: ; In the formula, d e Let q be the gate point. e For quality score, t now -t represents the difference in passage time between two consecutive access control events, and γ is the daily attenuation coefficient (e.g., 0.95). The standard address 'a' with the highest event score is selected and marked as the main address of the personnel file F; all event tuples in the personnel file F are sorted in ascending order of access time to form an access event sequence [e1, e2, ...]; from the sorted access event sequence, the door number of each access control event is extracted sequentially to obtain the door access sequence of that person [d1, d2, ...]; using the undirected door graph, the two adjacent door points d are queried. i With d i+1 The shortest path between each pair of adjacent gates is sequentially spliced ​​together to form a complete travel trajectory.

[0039] This invention's embodiment generates a verification work order based on the frequency of passage within a statistical window, including: If the personnel file is a stranger file and the number of times they arrive at any standard address within the long-term statistical window is greater than or equal to the high-frequency threshold, then an unregistered inflow verification work order will be generated. If the personnel file is matched with multiple passage times within a preset nighttime period after its first appearance in the short-term statistics window, an unregistered residence verification work order is generated. Within the short-term statistics window, the passage times of each passage event in the personnel file are counted, and target passage events with passage times within the preset nighttime period are filtered out. If the number of target passage events exceeds a preset frequency threshold, it is determined to be high-frequency nighttime passage, and an unregistered residence verification work order is generated accordingly. If the personnel file matches multiple non-adjacent standard addresses within a short statistical window, an abnormal wandering verification work order is generated. Within the short window, the corresponding door number (or standard address) in each access control event in the personnel file is extracted. If the door number corresponding to the door does not have a direct connected path in the preset door topology, it is determined that multiple non-adjacent door points have been matched, and an abnormal wandering verification work order is generated accordingly.

[0040] The long-term statistical window is, for example, 30 days or 60 days, and the short-term statistical window is, for example, 7 days. The activity status of personnel is confirmed through statistics from both the long-term and short-term statistical windows, and their address tags are updated accordingly, such as suspected residence status or residence status. For personnel who have been incorporated into the resident file in the previous steps, their address tag is residence status. For stranger files, by analyzing the access event sequence of each person, if the number of times they arrive at any standard address within the long-term statistical window is greater than or equal to the high-frequency threshold, or if they meet the high-frequency nighttime access pattern or hit multiple non-adjacent door points, an address tag is generated for the corresponding personnel file. The address tag includes suspected residence status and associated building information, thereby identifying suspected residents who have not appeared for a long time. The generated verification work orders (unregistered inflow verification work order, unregistered residence verification work order, abnormal wandering verification work order) remind the local grid administrator to conduct door-to-door verification to check whether they have moved out.

[0041] The access control event-driven dynamic personnel file management method described in this embodiment of the invention has the following advantages compared with the prior art: 1. This invention acquires access control events and facial images to obtain event tuples. The event tuples are compared with resident databases and stranger databases respectively. Through a judgment mechanism based on comprehensive similarity, dual-database matching, and merging confidence, the real-name merging path and the stranger merging path are separated. Finally, the resident files after real-name merging and trajectory aggregation are integrated with the stranger files to form a unified personnel file. This not only preserves the identity certainty of real-name personnel and avoids merging different people into one person, but also establishes traceable anonymous files for strangers through clustering technology. Ultimately, it achieves the personnel management goal of one file per person and full coverage.

[0042] 2. By calculating the gate topology and the theoretical shortest travel time, when the actual time interval is less than the theoretical shortest travel time, it is determined that there is a hard constraint conflict in travel time, and automatic merging is prohibited, which significantly improves the accuracy of file merging.

[0043] 3. Sort the event tuples in the same personnel file according to the time of passage, extract the door access sequence, and then calculate the shortest path between adjacent door points based on the door point topology and splice them in sequence to generate a complete passage trajectory. This allows property or security personnel to intuitively grasp the activity patterns of personnel and provides trajectory-level data support for abnormal behavior detection (such as loitering and tailgating).

[0044] 4. By analyzing long-term and short-term behavior, the system can automatically infer residential status without the need for manual door-to-door investigation. It can automatically mine hidden permanent residents from massive traffic data and associate suspected residential tags with standard addresses, providing accurate data clues for refined management.

[0045] 5. For all access control events, if the person is confirmed to be a resident, they will be added to the resident database; if the person is a stranger, a new database will be created or added to the stranger database. This achieves automatic registration upon arrival, eliminating the need for dedicated grid management personnel to conduct subsequent registration, and greatly improving the efficiency of registration and archiving.

[0046] Based on the above embodiments, as a supplement to the above... Figure 1 The present invention provides an embodiment of a dynamic personnel file management device driven by access control events, which implements the method shown. Figure 1 Corresponding to the method embodiments shown, this device can be specifically applied to various electronic devices, see reference. Figure 2 As shown, the access control event-driven dynamic personnel file management device includes: The event binding module 100 is used to obtain access control events reported by the access control controller, and to collect face images through the door point camera, and generate event tuples based on the access control events and face images; The constraint calculation module 200 is used to perform quality gating detection on the event tuple. If the quality gating detection is passed, the gate travel time is calculated based on the gate topology, and the comprehensive similarity is calculated by combining the facial feature similarity. The dual-database matching module 300 is used to perform dual-database matching on the resident database and the stranger database in descending order of priority based on the comprehensive similarity, and obtain the matching result; The file update module 400 is used to perform file creation or update operations based on the matching results. If no match is found with either the resident file in the resident database or the stranger file in the stranger database, a new stranger file is created based on the current event tuple. If a match is found with either the resident database or the stranger database, the file core of the corresponding file is updated based on quality weighting. The merge decision module 500 is used to calculate the merge confidence level when a stranger's profile matches a resident's profile in the resident database, and to perform automatic merging or generate a merge approval work order for approval based on a preset merge threshold strategy. The work order generation module 600 is used to generate a unified personnel file based on all merged resident files and stranger files, update the address tags corresponding to the personnel file and generate the passage trajectory, and generate a verification work order based on the passage frequency in the statistics window.

[0047] Each module in the aforementioned access control event-driven dynamic personnel file management device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0048] The access control event-driven dynamic personnel file management device described in this embodiment can execute the access control event-driven dynamic personnel file management method provided in the above embodiment. The access control event-driven dynamic personnel file management device has the corresponding functional steps and beneficial effects of the access control event-driven dynamic personnel file management method described in the above embodiment. For details, please refer to the embodiment of the access control event-driven dynamic personnel file management method described above. This embodiment will not be repeated here.

[0049] In one exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 3As shown, the computer device includes a processor, memory, input / output interfaces, a communication interface, a display unit, and an input device. The processor, memory, and input / output interfaces are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interfaces. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The input / output interfaces are used for exchanging information between the processor and external devices. The communication interface, such as a network interface card, is used for wired or wireless communication with external terminals. Wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. The display unit is used to create a visually visible image and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0050] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0051] In one exemplary embodiment, a chip is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of the method in any of the above embodiments.

[0052] In one exemplary embodiment, a network interface card is provided, including a chip as described in any of the above embodiments and multiple interfaces, wherein the chip communicates externally through the interfaces.

[0053] In one embodiment, a computer device is also provided, including a processor, a chip in any of the above embodiments, or a network interface card in any of the above embodiments, wherein the chip or the network interface card is used to schedule packets to the processor or the chip or the network interface card itself for processing, and the processor is used to process the packets scheduled by the chip or the network interface card.

[0054] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0055] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0056] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory may include read-only memory (Read-Only Memory). Only memory (ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory may include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in this application may include at least one of relational databases and non-relational databases. Non-relational databases may include distributed databases based on blockchain, etc., and are not limited thereto. The processors involved in the various embodiments provided in this application may be general-purpose processors, central processing units, graphics processors, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited thereto.

[0057] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0058] Similarly, it should be understood that, for the purpose of simplification and aiding understanding of one or more aspects of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof in the description of exemplary embodiments of the invention above. Other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the disclosure herein. This application is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only, and it should be noted that the above embodiments are illustrative of the invention and not restrictive, and that alternative embodiments can be devised by those skilled in the art without departing from its scope.

Claims

1. A method for dynamically managing personnel records driven by access pass events, characterized in that, include: S1. Obtain access control events reported by the access control controller and capture facial images through the door point camera, and generate event tuples based on the access control events and facial images; S2. Perform quality gating detection on the event tuple. If the quality gating detection is passed, calculate the gate travel time based on the gate topology and calculate the comprehensive similarity by combining the facial feature similarity. S3. Based on the comprehensive similarity, perform dual-database matching on the resident database and the stranger database in descending order of priority to obtain the matching results; S4. Perform a file creation or update operation based on the matching result. If no match is found with either the resident file in the resident database or the stranger file in the stranger database, a new stranger file is created based on the current event tuple. If a match is found with either the resident database or the stranger database, the file core of the corresponding file is updated based on quality weighting. S5. When a stranger's profile matches a resident's profile in the resident database, calculate the merge confidence level and execute automatic merging or generate a merge approval work order for approval based on the preset merge threshold strategy. S6. Based on all merged resident files and stranger files, generate a unified personnel file, update the address tags corresponding to the personnel file and generate a passage trajectory, and generate a verification work order based on the passage frequency in the statistics window; In step S2, the calculation of gate travel time based on gate topology and the calculation of comprehensive similarity based on facial feature similarity include: The gate topology is an undirected graph of gates. The shortest path walking time between two gates is calculated based on this undirected graph as the theoretical travel time. Determine whether the time difference between the current access control event and the historical access control event is less than the theoretical travel time minus the tolerance value; If so, the travel time is deemed unreachable, and the overall similarity score is set to invalid. If not, then the spatiotemporal affinity is calculated, which is determined based on the hierarchical relationship of the standard addresses corresponding to the current access control event and the historical access control events in the address tree; The comprehensive similarity is obtained by weighted summation of the spatiotemporal affinity and facial feature similarity.

2. The access control event-driven personnel profile dynamic management method according to claim 1, wherein, In step S1, the generated event tuple includes: Based on the access control events reported by the access controller, obtain the door number d, access time t, credential identifier c, and event type τ; Obtain the face image output by the gate camera and extract the face feature vector x from it; Based on the face image, the corresponding sharpness s, pose p, illumination l, and relative width z of the face frame are obtained respectively. The quality score q of the face image is calculated by weighted summation of the sharpness s, pose p, illumination l, and relative width z of the face frame. Map the gate number d to the standard address a; The event tuple e is generated based on the gate number d, passage time t, credential identifier c, facial feature vector x, quality score q, event type τ, and standard address a.

3. The access control event-driven personnel profile dynamic management method of claim 1, wherein, In step S3, the process of performing dual-database matching on the resident database and the stranger database in descending order of priority based on the comprehensive similarity includes: Confirm whether the credential identifier in the event tuple matches the credential field in the resident database. If the match is successful, bind the current event tuple to the corresponding matching resident file. If the credential identifier fails to match the credential field in the resident database, it is confirmed whether the overall similarity is greater than the resident similarity threshold corresponding to the resident file in the resident database. If it is greater than the resident similarity threshold, the current event tuple is merged with the corresponding resident file for real-name authentication. If the overall similarity is not greater than the resident similarity threshold, then it is confirmed whether the overall similarity is greater than the stranger similarity threshold corresponding to the stranger profile in the stranger database. If the overall similarity is greater than the stranger similarity threshold, then the current event tuple is merged into the corresponding stranger profile.

4. The access control event-driven personnel profile dynamic management method of claim 1, wherein, In step S5, when a real-name comparison between a stranger's profile and a resident profile in the resident database matches, the confidence level for merging is calculated, and automatic merging or approval merging is performed according to a preset merging threshold strategy, including: The merge confidence level is obtained by weighted summation of the average similarity between the stranger profile and the resident profile, the co-occurrence ratio on common gate points or adjacent addresses, and the real-name hit identifier. If the merge confidence level is greater than or equal to the automatic merge threshold, then the stranger profile and the resident profile are merged, retaining the primary key of the resident profile; If the confidence level of merging is less than the automatic merging threshold but greater than or equal to the approval threshold, a merging approval work order is generated, and the stranger's file and the resident's file are merged after the merging approval work order is approved.

5. The access control event-driven personnel profile dynamic management method of claim 1, wherein, In step S6, updating the address tag corresponding to the personnel file and generating the travel trajectory includes: For each of the aforementioned personnel files F, calculate the calculations at each standard address. a The event score (F, a) of the access control event is: ; In the formula, d e is the door point, q e is the mass fraction, t now -t is the difference between the access times corresponding to two consecutive access events, and γ is the daily attenuation coefficient. Select the standard address 'a' with the highest event score and mark it as the main address of the personnel file F; Sort all event tuples in the personnel file F in ascending order of passage time to form a passage event sequence; From the sorted access event sequence, extract the door number of each access control event in sequence to obtain the door access sequence of the person; Using an undirected graph of gate points, find the shortest path between two adjacent gate points, and then concatenate the shortest paths between each pair of adjacent gate points to form a complete travel trajectory.

6. The method for dynamic management of personnel files driven by access control events according to claim 5, characterized in that, In step S6, generating a verification work order based on the passage frequency within the statistical window includes: If the personnel file is a stranger file and the number of times they arrive at any standard address within the long-term statistical window is greater than or equal to the high-frequency threshold, then an unregistered inflow verification work order will be generated. If the personnel file appears for the first time within a short-term statistical window and is matched with multiple passage times within a preset nighttime period, an unregistered residence verification work order will be generated; If the personnel file matches multiple non-adjacent standard addresses within a short-term statistical window, an abnormal wandering verification work order is generated.

7. A personnel file dynamic management device driven by access control events, applied to the method described in any one of claims 1-6, characterized in that, The device includes: The event binding module is used to obtain access control events reported by the access control controller, and to collect face images through the door point camera, and generate event tuples based on the access control events and face images; The constraint calculation module is used to perform quality gating detection on the event tuple. If the quality gating detection is passed, the gate travel time is calculated based on the gate topology, and the comprehensive similarity is calculated by combining the facial feature similarity. The dual-database matching module is used to perform dual-database matching on the resident database and the stranger database in descending order of priority based on the comprehensive similarity, and obtain the matching result; The file update module is used to perform file creation or update operations based on the matching results. If no match is found with either the resident file in the resident database or the stranger file in the stranger database, a new stranger file is created based on the current event tuple. If a match is found with either the resident database or the stranger database, the file core of the corresponding file is updated based on quality weighting. The merge decision module is used to calculate the merge confidence level when a stranger's profile matches a resident's profile in the resident database, and to perform automatic merging or generate a merge approval work order for approval based on a preset merge threshold strategy. The work order generation module is used to generate a unified personnel file based on all merged resident files and stranger files, update the address tags corresponding to the personnel file and generate the passage trajectory, and generate a verification work order based on the passage frequency in the statistics window.

8. A computer device comprising a processor and a memory, characterized in that, The memory stores computer program instructions that can be executed by the processor, and when the processor executes the computer program instructions, it implements the instructions of the method as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer program instructions, which are loaded and executed by a processor to perform the operations performed by the method as described in any one of claims 1-6.